Tony Plummer - Forecasting Financial Markets The Psychology of Successful Investing 5th edition

In ForecasPlummer pinto the psbehaviour instinct canindividual imoney in man individuindependesuch indepacquire the

• understalogical te

• recognizemotionafluctuatio

• design atrading sobjective

The book edimensions

This fifth eFinancial Mupdated toauthor’s latthe implicaand rhythmfinancial mdeal with:

• the phen• the three• economi• recurren

activity;• forecasti• finding c

of is a

Bank Ltdment PLC.rket assets1999,currencies,dentd rhythms

wdalysis innd the Fargree fromterburythes. He is aechnicalactitionering.

“This book will entertain and intrigue keen investors.”

Financial Times

“Tony Plummer… has written a book that is almost breathlessin its enthusiasm for his chosen craft.”

The Independent

“A brilliant, original, insightful work… deserves to be readby all serious technical analysts.”

The Independent

“Should interest not only practitioners and skeptical economists, but alsocountless thoughtful managers who are rightly impatient with

expert ‘forecasts’, which are, alas, not always worththe paper they are printed on.”

Sir Adam Ridley (writing in the Foreword)

FORECASTING FINANCIAL MARKETSP

LUM

ME

R

5TH EDITION

Kogan Page120 Pentonville RoadLondon N1 9JNUnited Kingdomwww.kogan-page.co.uk

Kogan Page US525 South 4th Street, #241Philadephia PA 19147USA

Investment and securities

£45.00US $79.95

The Psychology of Successful Investing

FINANCIAL MARKETSFORECASTING

“This book will entertain and intrigue keen investors.” Financial Times

TONY PLUMMER

5TH EDITION

FORECASTINGFINANCIAL MARKETS

11/06 2:56 pm

7700++ DDVVDD’’ss FFOORR SSAALLEE && EEXXCCHHAANNGGEE

wwwwww..ttrraaddeerrss--ssooffttwwaarree..ccoomm

wwwwww..ffoorreexx--wwaarreezz..ccoomm

wwwwww..ttrraaddiinngg--ssooffttwwaarree--ccoolllleeccttiioonn..ccoomm

wwwwww..ttrraaddeessttaattiioonn--ddoowwnnllooaadd--ffrreeee..ccoomm

CCoonnttaaccttss

aannddrreeyybbbbrrvv@@ggmmaaiill..ccoomm aannddrreeyybbbbrrvv@@yyaannddeexx..rruu

SSkkyyppee:: aannddrreeyybbbbrrvv

http://www.traders-software.com/

http://www.forex-warez.com/

http://www.trading-software-collection.com/

http://www.tradestation-download-free.com/

mailto:[email protected]

mailto:[email protected]


i

ii

This page intentionally left blank


The Psychology of Successful Investing

5TH EDITION

TONY PLUMMER

London and Philadelphia

iii

To Glenys

First published in 1989Revised edition 1990Second edition 1993Third edition 1998Fourth edition 2003Fifth edition 2006

Apart from any fair dealing for the purposes of research or private study, or criticism orreview, as permitted under the Copyright, Designs and Patents Act 1988, this publicationmay only be reproduced, stored or transmitted, in any form or by any means, with the priorpermission in writing of the publishers, or in the case of reprographic reproduction, inaccordance with the terms and licences issued by the CLA. Enquiries concerning repro-duction outside those terms should be sent to the publishers at the undermentionedaddresses:

120 Pentonville Road 525 South 4th Street, #241London N1 9JN Philadelphia PA 19147United Kingdom USAwww.kogan-page.co.uk

© Tony Plummer, 1989, 1993, 1998, 2003, 2006

The right of Tony Plummer to be identified as the author of this work has been asserted byhim in accordance with the Copyright, Designs and Patents Act 1988.

ISBN 0 7494 4749 4

British Library Cataloguing in Publication Data

A CIP record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

Plummer, Tony.Forecasting financial markets : the psychology of successful investing / Tony Plummer. –

5th ed.p. cm.

Includes index.ISBN 0-7494-4749-41. Stock price forecasting. 2. Investment analysis. 3. Investments. I. Title.

HG4637.P57 2006332.63’2220112–dc22

2006013221

Typeset by Saxon Graphics Ltd, DerbyPrinted and bound in the United States by Thomson-Shore, Inc

iv

Contents

Foreword viiPreface to the fifth edition ixAcknowledgements xiii

Introduction 1

Part One: The logic of non-rational behaviour in financialmarkets

1 Wholly individual or indivisibly whole 92 Two’s a crowd 153 The individual in the crowd 264 The systems approach to crowd behaviour 335 Cycles in the crowd 416 Approaches to forecasting crowd behaviour 54

Part Two: The dynamics of the bull–bear cycle

7 The stock market crowd 738 The shape of the bull–bear cycle 869 Energy gaps and pro-trend shocks 107

10 The spiral and the golden ratio 12311 The mathematical basis of price movements 14012 The shape of things to come 153

Part Three: Forecasting turning points

13 The phenomenon of cycles 16514 The threefold nature of cycles 17615 Economic cycles 201

v

16 Recurrence in economic and financial activity 21717 Integrating the cycles 23218 Forecasting with cycles 24519 Finding cycles: a case study 25620 Price patterns in financial markets 27021 The Elliott wave principle 28622 Information shocks and corrections 30223 The confirmation of buy and sell signals 326

Part Four: The trader at work

24 The psychology of fear 35125 The troubled trader 36126 The psychology of success 37727 The mechanics of success 39328 Summary and conclusions 405

Index 409

vi Contents

Foreword

The last two decades have witnessed dramatic changes in thebehaviour of the free world’s financial markets. The fundamentalcauses of these changes must embrace both the end of fixed exchangerates in the early 1970s and the progressive removal of controls oninternational financial flows. However, whatever the precise reasonsmay be, the symptoms are evident:

� a very marked increase in the volatility of prices and volumes inmost markets;

� sharp and growing clashes between short-term developments andlong-term trends;

� striking contradictions between market sentiment and economicfundamentals.

The practical consequences have sometimes daunting and sometimeshumiliating challenges for forecasters but, far more importantly,much greater risk and uncertainty for businesspeople and traders.Despite the vast flow of data and major advances in computing andapplied statistics, conventional forecasting and economic analysisare all too often not providing the guidance market operators need.So, it is clear that we should now ponder on why this is and consider,in an open-minded way, what can be done to broaden and strengthenthe techniques we rely on. We need to widen our horizons so as to beable to heed the methods and results of technical analysis. Theminimum argument for doing so is the cynical or expedient one thatmany professionals in the financial markets draw heavily on tech-nical analysis in one way or another in their dealing as well as their

vii

writing and advising. A stronger argument is that technical analysisinvolves a serious attempt to reflect phenomena such as peer-grouppressure, fashion, crowd psychology and much else, which areignored or assumed unreliable by conventional theory. Regrettably,there has been little common ground sought between conventionalforecasters and technical analysts. Tony Plummer’s book is, in part, acontribution to that debate. However, it is also a serious attempt tostate systematically the basis of technical analysis in a way thatshould interest not only other practitioners and sceptical economists,but also countless thoughtful managers who are rightly impatientwith expert ‘forecasts’, which are, alas, not always worth the paperthey are printed on.

Sir Adam RidleyDirector General, London Investment Banking Association

Special Adviser to successive Chancellors of the Exchequer 1979–85

viii Foreword

ix

Preface to the fifthedition

We learn from our mistakes, and we interpret the world in the lightof our own experiences. Forecasting Financial Markets is a result ofboth these truths. In 1979, I was forced to learn that the economictheory that I had learnt at the University of Kent and the LondonSchool of Economics did not always work in the real world. At thetime, I was trading in the gilt-edged market in London. PrimeMinister James Callaghan had been overseeing a major inflationaryepisode, fuelled by rising government spending and acceleratingmonetary growth. Basic economics taught me that the UK needed asignificant tightening of monetary policy and that bond prices werein a serious bear market. However, in early 1979, gilt-edged stockprices rallied by about 20 per cent in the space of a few weeks,thereby retracing a good two-thirds of the previous year’s fall. Therally was triggered by what was subsequently referred to as the‘Battle of Watling Street’: representatives of City institutionsliterally fought to get into the Bank of England’s debt issuing officeprior to a 10 am deadline, in order to cover massive short positions.For a while, the civilized atmosphere of the City of London degen-erated into the physical behaviour of a crowd. Many missed thedeadline: demand for bonds exceeded the supply of bonds, andprices soared.

Extraordinarily, however, the rally was not just a normal, short-lived bear squeeze. It extended over eight weeks. Day after day, priceswent up, almost without pause. It was as if ‘fundamentals’ didn’t

matter; all that mattered was that investors were short of stock. Thenormally rational atmosphere of the London gilts market degeneratedinto the psychological behaviour of the herd. When Mrs Thatchercame to power in May 1979, monetary policy was tightened and giltprices spent almost a year losing the gains of February to May.Rational economics finally won the day, but a lot of investors lost a lotof money in the meantime.

This experience was a big lesson for me. The recognition thatpeople do not always act rationally, and do not always make decisionsindependently of one another, necessitated a shift in the way Iapproached financial markets. However, it was quite clear that therewas no central body of literature to which I could turn. So, within a fewweeks, I began my own process of collecting ideas and related infor-mation, and in 1989 – 10 years after ‘Watling Street’ – the resultingunderstandings found expression in the first edition of this book. Sincethen, there has been an explosion of literature from authors such asHoward Bloom (The Lucifer Principle, Atlantic Monthly Press, 1995,and Global Brain, John Wiley, 2000) that substantially validates myconclusions about the pervasive influence of crowd psychology.Economic man, who makes decisions on the basis of ‘rational expecta-tions’, is not only a travesty of a human being, but an impossibleconstruct. For good and for ill, human beings are literally programmedto coexist, cooperate and correlate with one another. Such relation-ships validate us and stimulate our emotional life.

This, in itself, is very important for our understanding of financialand economic behaviour. By definition, once people start to grouptogether, behaviour within the context of the group becomes non-random. This is why – despite what some statisticians may say –financial market price action has a non-random dimension. But this isnot all. Somewhat startlingly to someone who recognizes it for the firsttime, market price action persistently expresses itself in a three-wavepattern that mirrors the processes of learning, and of energy absorption,in living organisms. The crowd, in other words, is not different to otherparts of Nature. As such, it is intrinsically predictable.

Forecasting Financial Markets has now come a long way sincethose initial, exciting discoveries of the 1980s. For a start, myresearch revealed that the archetypal three-wave pattern existed notonly in financial markets, but also in economic activity. Moreover, Ifound that the nature of market oscillations pointed persistently to aspecific mathematical influence that allows us to identify the

x Preface to the fifth edition

difference between a technical reaction to an impulse wave and theemergence of a new trend. A wonderful, organized and essentiallypredictable world emerged from the apparent chaos. This fifth editionclarifies and expands on these findings. In particular, new sectionsdemonstrate how comparisons of cyclical patterns – both over timeand between markets – can yield extraordinary insights and conclu-sions. And the technique has been extended both to cover trends ininflation and to make some forecasts for the next decade. There aresome interesting – and worrying – conclusions.

Tony PlummerLittle Walden, Essex

Preface to the fifth edition xi

xii


Acknowledgements

Very few authors can have had their work published without the assis-tance and encouragement of other people. This one is certainly noexception.

My sincere thanks still go to Michael Hughes, who patiently readthe manuscript of the first edition and provided sound advice on howto recreate it in a more acceptable form. Needless to say, any subse-quent errors remain my own.

My thanks also to my former colleagues at Hambros Bank, espe-cially David Tapper and John Heywood. Their support and adviceover the years were major influences on my interest in the workings offinancial markets; and some of David’s wisdom in particular hasfound expression in the following pages, albeit in ways that he mightnot immediately recognize!

My thanks also to my wife and family. They accepted mycommitment and provided complete support for the task, even when Ipersistently spent long evenings and weekends hunched over a wordprocessor.

Finally, my thanks must also go to my travelling companions on the7.30 am from Audley End. It was their dedication to either sleep orcrossword puzzles that enabled the earliest version of the manuscriptto be written in the first place.

xiii

xiv


Introduction

Making money by trading in financial markets is a formidable task.This is a great truth that is almost impossible for one person to teach toanother. It can only be realized by the very act of trading. Accordingly,very few people enter the trading arena with their eyes fully open tothe psychological and financial risks. Indeed, they approach marketsin the same way that they might approach a lake containing a fabledtreasure. They feel that all they have to do is set up appropriatepumping equipment and the treasure is theirs. What they do not realizeis that – as in all good fairy stories – the lake has magical properties,designed to protect the integrity of the treasure. Most people whotouch the sparkling water of the lake are doomed to be transformed byit: they become treasure protectors instead of treasure hunters.

The problem is that there is an energy in financial markets (and,indeed, in economic activity) that somehow coerces and organizesinvestors into a single-minded unit. There is nothing sinister in this: itis just nature ‘doing its thing’. However, the force is a psychologicalone, and it is so powerful that investors do not recognize it until theyare finally caught in a disastrous bear market that wipes out months, ifnot years, of hard work.

The corollary of this is that truly great traders are very rare – only afew have the special clothing that protects them from the secret effectsof the lake. Jack Schwager’s books on ‘Market Wizards’ would nototherwise be best-sellers.1 What is it, though, that separates suchtraders from the rest of us? How can some traders make regular andlarge profits, while others are unable to string two successive winningtrades together? The traditional answer is that the quality of thetrading system may not be good enough. Much time and money istherefore spent on developing trading systems in order to generate

1

improved entry and exit signals. The problem, however, is that a goodtrading system – while absolutely essential – is only part of thesolution. It is also necessary to be able to implement the signals on aconsistent basis.

This may seem like a trivial statement, but it is not.2 The reality isthat successful trading requires a certain psychological competence todo the job and, unfortunately, nature has chosen not to wire up thehuman psyche with automatic access to this competence. The pointhere is that financial markets can have a direct and dramatic effect onwealth and on associated living standards. Most people are thereforelikely to experience an emotional response as the market adds to, orsubtracts from, the value of their assets. This is not necessarily bad initself, but it does mean that when a system generates a buy or sellsignal, the investor will still feel obliged to decide whether or not toimplement that signal. Investment thereby moves away from theobjective realm and into the subjective realm.

On this analysis, success or failure hinges critically on the ability topenetrate the ‘emotional gateway’ between the generation of a ‘buy’or ‘sell’ signal and the implementation of that signal. Penetrating thisgateway is, in essence, what this book is about. However, success isnot just a case of heavy armour and battering rams. Such methods canwork for short periods of time, but they require a heavy expenditure ofenergy and, inevitably, exhaustion sets in. A much more rewardingapproach is to see the gateway as a learning opportunity. For the truthis that, once the necessary learnings have been completed, thegateway simply disappears.

Nor are the learnings particularly difficult. We are not talking abouta lifetime of struggle that then enables a magnificent and heroicinvestment decision to be made from the deathbed. We are talkingabout relatively small adjustments in understandings, attitudes andresponses, so that a large number of successful decisions can be madeon a continuous basis. In fact, it is not too idealistic to say that we aretalking about a commitment to the truth of the situation. Such acommitment encourages flexibility, without which we cannot respondeffectively to unexpected information: that is, we cannot admit to ourmistakes, we cannot be free of other people’s opinions and we cannotmake appropriate decisions.

Ultimately, it doesn’t matter what economic theory predicts or whatyou believe ought to happen or what other people believe will happen.All that matters is that you have aligned yourself with the actualmarket trend, and if this involves changing your investment position,your belief system or even your drinking companions, then so be it.

2 Forecasting financial markets

The argument of this book, then, is that flexibility in the decision-making process can be attained by generating three interrelated skills.The first of these is an ability to understand the market in logicalterms. That is, an investor or trader should have a philosophicalapproach to markets that incorporates a genuine understanding of theforces at work. The point is that markets fluctuate – regularly and,according to traditional theory, unpredictably. If an investor does notclearly recognize this, then they will be unprepared for the reality – theterror – of the situation.

The second skill, in a sense, follows from this. An investor needs tobe able to understand their own emotional response to market fluctua-tions. This is a part of what Daniel Goleman calls ‘emotional intelli-gence’.3 If market participants understand their own vulnerability tothe influences of financial markets, and if they can recognize thoseassociated behaviours that are potentially self-defeating, then they cando something about it. In particular, they can adopt responses that areappropriate to market trends rather than responses that are hostile tothem. The alternative, quite simply, is to become a victim.

Finally, of course, an investor needs to be able to design aninvestment process, or trading system, that generates objective ‘buy’and ‘sell’ signals. In other words, the signals must be based on pre-determined criteria. This approach has two advantages:

� first, it focuses attention on critical factors that tend to recurthrough time;

� second, it helps to reduce the influence of any emotions that occurat that point in time.

Such a system need not be mechanical: it can have a facility to incor-porate signals that cope with unusual circumstances or with investorpreferences. The only criterion that ultimately matters is that theinvestor/trader is sufficiently confident about the signals that they willnot continuously be doubted. Market participants should, therefore, bedirectly involved in the system testing procedures, so that they areaware of both the successes and the limitations. This, of course, alsofacilitates a creative response to unexpected market developments.

These three skills – the ability to understand market behaviour inlogical terms, the ability to know the effects of the market inemotional terms and the ability to decide what to do in objective terms– are the basis of successful wealth creation in financial markets.Further, they enable investors and traders to be detached from theresults of each individual investment position in a way that enables

Introduction 3

them, literally, to enjoy the whole process. In this way, wealth creationand personal fulfilment become a way of life.

This book explores each of these three dimensions to successfultrading in some detail. The first step is to analyse the phenomenon ofthe ‘crowd’ in financial markets. Crowds come into being because ofthe existence of common beliefs and because of the need forprotection from opposing beliefs. It is the dynamics of the crowd thatcause markets to fluctuate as they do. Over the course of apronounced market trend, investors become increasingly unable topenetrate the emotional gateway that we described above. Instead,they tend to do the same things at the same time in order to obtainpsychological support. Eventually, a significant price reversal occurssimply because the majority are effectively ‘one-way’. By definition,therefore, most investors will be on the wrong foot when a reversaloccurs. It is this fact, more than anything else, that explains traumassuch as the stock market crash of 1987 and the bond market crash of1994. Once prices start to accelerate in the opposite direction,investors and short-term traders are induced to close off old positionsand open new ones. Such behaviour maximizes losses and minimizesprofits. The result is that very few investors make consistently highreturns over and above bank deposit rates or outperform theirbenchmark indices.

The idea that people have a tendency to herd together is, of course,not a new one. What is not yet clearly understood, however, is thatgroup, or crowd, behaviour is an unavoidable feature of the humancondition. The crowd is a potent force because it encourages indi-viduals to subsume their own needs to those of others. This trans-ference of responsibility introduces a very large non-rational, andemotional, element to behaviour. Of course, very few people actuallyrecognize the influence of the crowd because, as the saying goes, ‘fishdon’t know that they’re swimming in water’. However, the balefulinfluence of the crowd permeates all economic and financialbehaviour, and is particularly noticeable in financial markets.

This means that it takes only a small adjustment in our assumptionsconcerning the nature of human motivation to generate a huge leap in ourunderstanding of observed human behaviour. Specifically, if we acceptthe assumption that individual behaviour is influenced to some degree bythe need to associate with – and obtain the approval of – other people,then all economic and financial behaviour can be seen as being orderedrather than chaotic. The uncertain behaviour of the individual transmutesinto the more certain behaviour of the crowd. As a result, economic andfinancial activity become more explicable and predictable.


Importantly, though, this predictability is inherent in the processitself. For example, it is not necessary to look at outside forces toforecast the rhythm of breathing; once one cycle has been measured, itis possible to forecast the next cycle. This is the essence of the ideathat market trends can be forecast with a high degree of accuracy byfocusing attention, not so much on external economic trends andvalues (although these are important), but on what other investorsthemselves are actually doing about these economic values. In thisway, it is possible to arrive at decisions about the position of themarket within the context of its trend, and it is possible to make asso-ciated investment decisions that are relatively uncontaminated by thepressure to conform to group beliefs.

The idea that market trends can be anticipated by analysing theactual activity of investors is the central tenet of the trading disciplineknown as ‘technical analysis’. Such analysis assumes that all financialmarkets follow specific behavioural laws, the influence of which canbe observed in price–time charts and in associated indicators ofinvestor activity, such as volumes and momentum. Over the years, awide range of trading techniques has been developed to take advantageof these laws. However, the problem has always been to explain whythe laws exist in the first place. Technical analysts might be prepared toassume that the laws are a fact of (and a gift from) nature, but others –especially academic economists – have tended to dismiss the laws asbeing ‘accidental’ and in conflict with common sense. While it wouldbe an arrogance to claim that the analysis contained in this book iscomplete, it nevertheless demonstrates that the phenomenon of thecrowd is a justifiable theoretical basis for technical analysis.

Importantly, many of the ideas that are involved in demonstratingthis point can be explored using the language of natural systemstheory. This discipline was originally developed to deal withbiological phenomena, but has since been found to have wider impli-cations. The exploration will necessarily be highly simplistic, but itwill be demonstrated that natural processes create a specific (andcontinuously recurring) price pattern, which is essentially the blue-print for market movements. For ease of exposition, we have calledthis pattern ‘the price pulse’.

Despite the apparent novelty of the price pulse, it must be empha-sized that it does not supplant other techniques. Indeed, it can beargued that not only does the price pulse completely validate tradi-tional technical analysis (which incorporates phenomena such as‘trend lines’ and ‘head-and-shoulders’ reversal patterns), but that it isalso the basis of the important ‘Elliott wave principle’.

Introduction 5

It will be shown that the price pulse is subject both to simple math-ematical relationships between its constituent phases and to regularrhythmic oscillations. Hence, it is possible to create a workable, andoften uncannily accurate, ‘map’ of likely future movements. Suchmaps are able to show the likely:

� profile of price movements;

� extent of those price movements;

� timings of price reversals.

Furthermore, it will be demonstrated that expected price reversals canbe confirmed in real time by direct reference to certain simplemeasures of investor behaviour. Price maps can therefore be used aspart of an effective trading system.

An intellectual understanding of the forces at work in financialmarkets, and the creation and use of a structured decision-makingprocess, provide an extraordinarily powerful basis for tradingmarkets. As already indicated, however, intellectual and technicalrigour are only part of the challenge. Truly successful investmentrequires a certain degree of psychological competence. Without this,even a system with 20/20 foresight might be doomed to failure. Asignificant part of this book is therefore devoted to explaining howsuch psychological competence can be attained. Unfortunately, abook – any book – cannot ‘teach’ psychological competence: it canonly point the way. That way is not easy for the vast majority of us; butthe effort can bring untold rewards.

The overall aim of this book, then, is to enable the investor tounderstand the dynamics of financial markets, to understand the asso-ciated personal competencies that are required for investing in themand the essential format for an objective decision-making system.

The only way, however, to know whether or not these guidelinesactually work is for you to try them. Good luck.

NOTES

1. Schwager, Jack (1992) Market Wizards, Wiley, New York.2. Indeed, in a sense, all great truths are essentially simple to comprehend.

Complexity is a characteristic of the periphery of a system; simplicity is a char-acteristic of its centre.

3. Goleman, Daniel (1996) Emotional Intelligence, Bloomsbury, London.


7

Part One

The logic of non-rationalbehaviour in financialmarkets

8


9

1

Wholly individual orindivisibly whole

INTRODUCTION

Western culture places a great deal of emphasis on the rights of theindividual. The concept of freedom of expression and the right of self-determination are enshrined in the democratic political systems ofNorth America and Western Europe. Indeed, they are so familiar to usthat most of us do not give them a second thought. Nevertheless,single-minded concentration on the needs and desires of the indi-vidual has encouraged an arrogance that is at once both an asset and aliability. It is an asset because it has catapulted humankind on avoyage of discovery through the universe that is within each of us, butit is also a liability because it has encouraged us to place ourselvesabove the cosmos of which we are a part.

Many scientists and philosophers now believe that future progresswill depend on our ability to recognize and accept that the inde-pendence of each individual is a relative condition rather than anabsolute one. Humankind takes great pride in its control and directionof certain aspects of the environment, but it still remains true that ulti-mately we are all dependent on that environment in the crucial senseof being a part of it. In fact, one of the most exciting features of scien-tific research during the last 50 years is the recognition that there is adeep interrelatedness among natural phenomena. Quite simply, every-thing in nature depends on everything else.

THE RELATIONSHIPS IN NATURE

This finding has significant implications for the development ofhuman knowledge, because it suggests that the most important aspectof the world is not the individual parts of nature so much but the rela-tionships in nature: the relationships define the parts, and no singlepart can exist independently of other parts.1 Hence, it becomespossible to visualize the world in terms of multilevel structures thatstart at the subatomic level and then extend upwards in ever-increasing layers of complexity. As an example, electrons combine toform atoms, atoms combine to form molecules, molecules combine toform organs, organs combine to form organ systems, organ systemscombine to form animals and humankind.

THE BREAK WITH TRADITION

These concepts have been explored in some detail in recent years,2and have even given a strong impetus to a new discipline known as‘systems theory’.3 However, the ideas are not yet widely understood.Part of the difficulty derives from the fact that systems theory marks adistinct break from the traditional analytical procedures that havebeen favoured ever since the pioneering work of Isaac Newton andRené Descartes. These procedures presume that it is possible to under-stand all aspects of any complicated phenomenon by ‘reducing’ thatphenomenon to its constituent parts.

The process of dividing nature into progressively smaller units (aprocess that is known as ‘reductionism’) works very well in thecontext of everyday life. Indeed, the fund of knowledge is actuallyenhanced as differentiation increases, and so the process is self-justi-fying. However, in the 1920s, physicists found that the process wastotally inapplicable at the subatomic level.4 Specifically, it was foundthat electrons do not exist with certainty at definite places and do notbehave predictably at definite times.5 In other words, there was acritical level where ‘certainties’disappeared, and where the concept ofbasic ‘building blocks’ seemed to become invalid.

The practical solution to the problem was to step back and assigncharacteristics to electrons that accounted for both the uncertainty ofthe unobserved state of existence and for the certainty of the observedstate. It was hypothesized that electrons had a dual nature: on the onehand, the behaviour of an individual electron could not be forecast

10 The logic of non-rational behaviour in financial markets

with any degree of certainty; on the other hand, the behaviour ofgroups of electrons could be forecast with a high degree of certainty.In other words, the solution6 lay within the mathematics of probabilitytheory, where a large number of uncertainties produce a certainty.7Probability theory can, for example, determine with 100 per centaccuracy the half-life of any radioactive substance,8 despite the factthat the point in time when one particular radioactive atom will disin-tegrate is totally unpredictable.

THE CONCEPTUAL REVOLUTION

The search for basic building blocks in nature will undoubtedlypersist.9 In the meantime, the revelations of the New Physics (as it iscalled despite the fact that it is well over half a century old) are gener-ating major structural changes in the natural and social sciences.10

Each discipline, essentially, is having to absorb two related ideas:

� ‘wholes’ are something greater than the simple arithmetic sum oftheir ‘parts’;

� each ‘part’ has a tendency to have both a separate identity and tobelong to a greater whole.

THE PROBLEM OF MOTIVATION

In the social sciences, the changes are leading towards a revolution inour understanding of human behaviour. In economics, for example,the traditional approach has been to assume that human beings areessentially mechanistic in their behaviour patterns. This assumptionis based partly on introspection and partly on research. First, even anelementary level of self-analysis reveals that a large part of ourbehaviour involves an automatic response to particular stimuli.Second, statistical analysis can also be used to show that groups ofpeople tend to respond in a predictable way to given stimuli. It is,therefore, only a small step to infer that individual behaviour is defi-nitely mechanistic and to use that inference as the basis foreconomics analysis and forecasting.

It is certainly true that a great deal of our behaviour is mechanistic.This is, in part, due to fundamental biological and ‘social’ drives that

Wholly individual or indivisibly whole 11

are common to many species that inhabit this planet. Indeed, anthro-pologists such as Desmond Morris11 have been able to identify a largenumber of parallels between human behaviour and that of animals.However, mechanistic behaviour is also due to a type of habitformation that is a particular feature of the human subconscious mind.The point is that, because individual histories differ, everyone’s habitsare going to be different to some degree, no matter how small. This, ofcourse, leaves an inconsistency between habitual personal behaviourand predictable group behaviour.

There is, however, a more telling criticism. This is that humanbeings have a type of awareness that specifically transcends automaticbehaviour. As Fritz Schumacher12 observed, the structure of livingorganisms is a progression of increasing complexity and power: plantshave life; animals have life and consciousness; people have life,consciousness and self-awareness. Self-awareness here means theability to be conscious of one’s own existence,13 and it encourageseach individual to choose between alternative responses to a givensituation. Self-aware decisions are not forecastable by outsiders. Noris there any reason to believe that one person’s decisions will be thesame as anybody else’s. As a very simple example, if someone wasgiven a windfall sum of money, it would be very difficult for anotherperson to predict precisely how that money would be used: it could besaved, spent, lent, given away or even destroyed.

THE DUALISTIC NATURE OF MOTIVATION

Economists have tended to ignore the inconsistency between indi-vidual variety and group conformity for two reasons. First, it is oftenargued that the assumptions in a theory are less important than theconclusions of that theory.14 Although there are some circumstanceswhere this may be true,15 it is by no means always correct. Inparticular, assumptions are all too often reflections of biased beliefsabout the world. Second, and perhaps more importantly, economictheoreticians have tended to resist the idea that a paradox actuallyexists. The reason for this seems to be that it presents a very real threatto the logical structure of modern economic theory (this point iscovered in more detail in Chapter 6).

The solution to the dichotomy between individual variety and groupconformity lies in the concept of a duality of characteristics, compa-rable to that used for subatomic phenomena. People have both the


ability to be individuals and the tendency to belong to groups. Theactual mix of the two characteristics varies over time depending oncircumstances. Sometimes a person will be relatively individualistic,while at other times the same person will be relatively willing toconform to behavioural patterns pursued or imposed by others. Theimportant difference between the two sets of circumstances is thedegree to which a person accepts other people’s belief systems,thereby limiting their personal room for manoeuvre.

CONCLUSION

The idea that motivation has a dual nature represents a major break-through in our understanding of human behaviour. Each person simul-taneously has a tendency to be an individual with a unique and‘personal’ view of the world and a tendency to belong to groups.Individual behaviour is not easily predictable, but group behaviour is.This duality of character is of the utmost importance to our analysis ofinvestor behaviour in stock markets.

NOTES

1. The interconnectedness of nature is actually such that it can sometimes be verydifficult to see where each ‘part’ of a structure starts or where each ‘whole’ends. See Bohm, David (1980) Wholeness and the Implicate Order, Routledge& Kegan Paul, London.

2. See, for example, Jantsch, Erich (1980) The Self-organizing Universe,Pergamon, Oxford.

3. See Chapter 4.4. The findings of the New Physics are truly remarkable. It is suggested that inter-

ested readers refer to, for example, Zukav, Gary (1979) The Dancing Wu LiMasters, William Morrow, New York.

5. Electrons can be viewed as being very abstract packets of energy that have adual aspect: sometimes they adopt the characteristics of a single entity orparticle, but sometimes they adopt the characteristics of continuous waves.Hence, prior to observation, it is impossible to determine whether an electron isparticle-like or wave-like. Furthermore, during observation, it is impossible todetermine both the position and the velocity of an electron.

6. Heisenberg, Werner (1971) Physics and Beyond, Allen & Unwin, London.7. This point is well made by Arthur Koestler (1978) in Janus: A summing up,

Hutchinson, London.8. The half-life of a substance is the time required for half the atoms in the

substance to disintegrate.

Wholly individual or indivisibly whole 13

9. The search for basic building blocks of matter has recently focused on smallerphenomena called ‘quarks’. However, even ignoring the difficulties regardingthe status of electrons, it is not certain that the concept will be validated. See,for example, Hawking, Stephen W (1988) A Brief History of Time, TransworldPublishers, London.

10. Capra, Fritjof (1982) The Turning Point, Wildwood House, London.11. Morris, Desmond (1967) The Naked Ape, Jonathan Cape, London.12. Schumacher, E F (1977) A Guide for the Perplexed, Jonathan Cape, London.13. The term ‘self-awareness’ can be used as a generic term, which incorporates all

the factors that enable the human mind to create an inner world that mirrors theouter ‘reality’.

14. See, for example, Friedman, Milton (1953) Essays in Positive Economics,University of Chicago Press, Chicago.

15. The argument that assumptions are not as important as the conclusions (orpredictions) of a model has some validity under two circumstances. It isacceptable when the assumptions are designed to neutralize external interfer-ences. In this way, it becomes possible to analyse the response of specific vari-ables to changes in a limited number of other variables, with outside influencesbeing held unchanged. The argument is also acceptable when the assumptionssimply affirm the ‘truth’ of previous research. Serious philosophical problemsarise, however, when the assumptions are used to neutralize complicationswithin a particular system because, by definition, these complications are partof the system.


15

2

Two’s a crowd

INTRODUCTION

Investors who have been involved with financial markets for anylength of time will readily identify with the concept of a conflictingtwo-way pull on their decisions. On the one hand, their own ‘personal’approach to making an investment decision may suggest one course ofaction; on the other, the lure of the ‘herd instinct’ may be pullingentirely in the opposite direction. Even seasoned professionals, whomake a living by anticipating and outwitting the rest of the partici-pants in financial markets, will sometimes find themselves caught upin a common hysteria at just that time when contrary thinking is trulyappropriate. It is a rare trader who can honestly say that they have notbought stock at the top of a price movement or sold stock at the bottomof a price movement.

This two-way pull on each individual is actually a natural result ofthe intrinsic relationship between individual integrity and groupmembership. On the one hand, each person has a self-assertivetendency or ability to behave in a self-determined, individualistic way.On the other hand, however, each individual also has an integrativetendency,1 which involves a psychological drive to belong to groupsof all kinds. Membership of a group involves the adoption of behav-ioural patterns that are consistent with the group’s objectives.


THE INFLUENCE OF GROUPS

It is important to recognize that group behaviour is not, in itself,unusual. In fact, groups emerge as a result of the same basic laws thatapply to the rest of nature. As Erich Jantsch2 has shown, building onthe implications of quantum physics, all of nature consists of multi-levelled structures. Each level in this hierarchy has the power toorganize its lower levels and use them for its own purposes.3

Consequently, each level is able to perpetuate itself or maintain itsidentity despite changes in its individual components.4 This hierar-chical structure applies to human society:5 individuals becomemembers of groups, groups merge to form societies and societiesmerge to form civilizations (see Figure 2.1).

THE INSIGHTS OF GUSTAVE LE BON

One of the first people to analyse the phenomenon of human groups inany detail was Gustave Le Bon.6 He was fascinated by the influenceand role of a very particular type of group – namely, the crowd – in theunfolding of the French Revolution. His seminal book ThePsychology of Crowds was written in 1895, but it still stands out as aclassic of social psychology. Its analysis has been validated by subse-quent analysts such as Sigmund Freud and Carl Jung, as well as bytheorists such as Arthur Koestler. Its conclusions have been found tobe applicable to such diverse historical phenomena as the NurembergRallies and the Holocaust in Nazi Germany; the attempted destruction

Figure 2.1 The hierarchical structure of human groupings

Two’s a crowd 17

of ‘bourgeois’ and reactionary values during the Cultural Revolutionin China; and the attempted elimination of individualism by theKhmer Rouge in the Killing Fields of Cambodia.

Le Bon saw a crowd as being primarily a psychologicalphenomenon rather than a physical one (although the two concepts arenot necessarily mutually exclusive). He considered that any number ofotherwise independent and spatially separate individuals could form acrowd, provided that the members had a common cause. This, ofcourse, confirms the idea that ‘crowd’-type pressures can be found ina large range of groupings. They can, for example, be found incompanies, football teams, armies, religious sects and patriotic nationstates, as well as riotous mobs. As we shall see, they can also be foundin financial markets. Le Bon himself argued that:7

The most striking peculiarity presented by a psychological crowd is thefollowing: whoever be the individuals that compose it, however like orunlike their mode of life, their occupations, their character, or their intel-ligence, the fact that they have been transformed into a crowd puts themin possession of a sort of collective mind that makes them feel, think,and act in a manner quite different from that in which each individual ofthem would feel, think, and act were he in a state of isolation . . .

What really takes place [in the formation of a crowd] is a combi-nation followed by the creation of new characteristics, just as in chem-istry certain elements, when brought into contact . . . combine to form anew body possessing properties quite different from those of the bodiesthat have served to form it.

This profound insight into the nature of crowds used two importantconcepts:� a crowd is something other than the sum of its parts – in particular,

a crowd has an effective ‘mind’ of its own;� each individual’s behaviour is altered by their membership of a

crowd.

These two concepts are now central to the general theory of groupbehaviour. Within this theory, the phenomenon of the crowd presentsitself as a special, and extreme, case.

THE GROUP’S ABILITY TO ORGANIZE ITSELF

From the group’s point of view, it is crucial to its own autonomy that itcan ‘organize’ its own membership. This is the only way that energy

can be directed towards the overall objectives of the group. Such‘self’-organization operates through the group’s belief system. To be amember of a group, an individual has to accept uncritically the samebeliefs as other members of the group. In 1961, the psychoanalystWilfred Bion proposed that these beliefs were ‘basic assumptions’held by group members.8 In 1976, Richard Dawkins called thebeliefs/assumptions ‘memes’.9 A meme is a simple, self-replicatingidea (or habit, or feeling, or sense of things) that spreads from mind tomind. The presence of a particular meme does not necessarily excludeother memes, so a group member can belong to more than one group.However, whatever term might be used for the commonly heldbeliefs, group members will necessarily suppress behaviour thatmight cause them to be excluded from the group.10 In other words, aperson’s individuality – or self-assertiveness – will be modified bygroup membership.

This situation obviously means two things:

� the more intense is the integrative pull of a group, the less is thedegree of individuality that is allowed (we shall deal with this inChapter 3);

� it is the like-mindedness of group members that gives the group itscohesion.

A group therefore organizes its members via a unifying belief system.In this way, a group becomes something more than just the sum of itsparts and group members will reveal a sense of altruism towards oneanother in the context of group activity. Hence – whether they overtlyrecognize it or not – group members will act in the interests of thegroup as a whole.

MIND AS A DYNAMIC PRINCIPLE

This analysis has necessarily been very brief and hardly does justice tothe concepts involved. Nevertheless, we now have the tools to enableus to look a little more closely at the idea of a group ‘mind’. This ideais not an easy one either to grasp or to convey. Part of the difficulty liesin our language itself. As commonly used, the word ‘mind’ is taken torefer to that part of the physical structure of the brain that is capable ofself-aware, rational thought. According to this view, people haveminds, but animals do not. However, this usage not only ignores therole of the so-called ‘subconscious’ mind, but assumes that the


physical structure of the brain is identical to, and provides the defininglimits for, the inner processes of the mind.

The best way to understand the difference between the concept of‘brain’ and the concept of ‘mind’ is to recognize that the brain is thestructure, while the mind is the processing capabilities that are origi-nated within that structure. Importantly, however, the processes of themind are not confined to the boundaries established by the physicalbrain: laterally, the processes extend through the living body, dealingwith automatic functions, physical movement and emotions; and,vertically, they extend (layer upon layer) into the depths of the person-ality. These processes are known generically as ‘mentation’.

The essential point to grasp at this stage is that mentation is prima-facie evidence of the phenomenon of life. This was the profoundinsight of the great philosopher and biologist Gregory Bateson.11

Bateson found that exactly the same characteristics that definementation in the human brain can be found both in all the otherprocesses of the human body and in every aspect of nature. Mentationis actually a logical process – a single dynamic blueprint – that is thevery hallmark of life on this planet. It does not, therefore, have to beencased in any particular physical structure and, conceptually, canextend beyond all physical structures. The difference betweenmentation in the human mind and mentation in other aspects of natureis in the quality, or depth, of consciousness. Only human mentationallows for thinking to be aware of itself.12

Bateson’s own criteria for the existence of mind (that is, the exis-tence of ‘life’) in any system were essentially four-fold, namely:

� the ability to control functions that were internal to the system;

� the ability of the system to process information;

� the presence of fluctuations in the processing of that information;

� the existence of a continuous exchange of energy and informationbetween the system and its environment.

As we shall see, these criteria produce some very important insightsinto human behaviour.

THE GROUP MIND

Bateson’s criteria for the existence of mind extended the concept toinclude all aspects of the living universe and, it may be said, added an

Two’s a crowd 19

important new dimension both to the conclusions of the New Physicsand to the framework of the scientific process. Hence, the concept ofmind can be used not only to describe the phenomenon of life, but also todescribe and explain any particular ‘unit’ of life. It can be found in boththe most simple and the most complicated of processes. In terms ofhuman society, it can be found in the dynamics of people’s relationshipswith one another: it can be found in a small group of people, in a physicalcrowd, in a nation state or in a whole culture. An individual’s mind may(in a non-self-assertive state) be regarded as a subsystem of a greaterwhole. Each whole, in turn, has a ‘collective mind’that organizes its ownparts. From this analysis we could say that each and every humangrouping may be said to have a collective consciousness.13

THE TRIUNE HUMAN BRAINAs already observed, however, the collective consciousness of ahuman grouping is qualitatively different from the consciousness ofan individual human being. The former is essentially simplistic and isnot usually recognized, while the latter is not fully understood, eventhough it is accepted. The essential difference between the twophenomena is that a group does not have the ability to be aware of itsown existence, whereas an individual does have such an ability. And itis this difference that is so important to our study of the influence of acrowd. The fact is that, as described by Le Bon, crowd behaviourinvolves a dramatic suppression of self-awareness by individuals andan increase in the power of the group entity that is, by definition, notself-aware in the first place. The ‘total’ quality of consciousnesstherefore deteriorates. As Scott Peck put it, ‘groups are, from apsychological standpoint, less than the sum of the parts’.14

We can isolate two aspects to this important phenomenon. First,when an individual adopts a group belief, then, by definition, they alsoaccept a reduction in self-awareness. This is so because the bound-aries between the individual and the group become blurred. However,a reduction in self-awareness necessarily happens in all groupbehaviour. What makes crowd behaviour so distinctive is the degreeto which an individual’s self-awareness is suppressed. It seems almostas if there is a switch somewhere in the brain that flicks self-awarenessto ‘off’ whenever a crisis arises.

This brings us to the second aspect of the deterioration in thequality of consciousness that accompanies the crowd phenomenon.There does seem to be something of a fault-line within the structure


of the human brain. The problem stems from the fact that the brainconsists of three main processing areas that can, to some extent,operate independently of one another. Each part is structurally andchemically different from the other parts, and appears to have itsown intelligence, its own memory and its own separate functions.The brain stem (the innermost part of the brain) is concernedprimarily with instinctive behaviour patterns, biological drives andcompulsive behaviour. Surrounding this part is the limbic system,which is mainly involved with the recreation of external experiencesin the ‘inner’ world, and with emotional activity. Importantly, it isalso the area of the brain that is most concerned with group activ-ities. Finally, the neocortex (the outermost part of the brain) dealswith the ability both to be aware of the thought process itself, and toanticipate the future and recreate the past. It also contains the areasof the brain (the so-called ‘frontal lobes’) that are most involvedwith concern for humanity.15,16 Because of its three-fold nature, theAmerican neurophysiologist Paul D Maclean17 has called the humanbrain the ‘triune’ brain.

THE NEOCORTEXOf the three areas, it is undoubtedly the neocortex that (among otherthings) separates humankind from other mammals. The evolutionaryhistory of the neocortex is, however, uncertain. No one really knowswhen (or how, or why) it first appeared. Current guesswork suggests itdeveloped only during the last 50 million years or so. This compareswith more than 150 million years for the limbic system, and more than250 million years for the brain stem. Importantly, evidence from theevolution of civilization18 suggests that the specific abilities asso-ciated with the neocortex have only developed within a very recentperiod. The modern ego – that is, the tendency of mental processes toorganize themselves in such a way as to assert themselves as an inde-pendent entity – seems only to have arisen some four thousand (or so)years ago. The ego, therefore, is very young in terms of evolution.

We can therefore hypothesize that the operation of the neocortexcan easily be overwhelmed for two related reasons:

� the neocortex has still not yet been properly integrated with theother two parts of the brain;

� the capability of ‘self’-awareness is still likely to give way to morearchaic states of consciousness when an ‘external’ authority isfavoured.

Two’s a crowd 21

THE AMYGDALA

In fact, research initiated by Joseph LeDoux at New York Universityhas found that the main problem almost certainly rests with a tiny partof the human brain known as the ‘amygdala’.19 The amygdala is asmall neural cluster lying between the brain stem and the limbicsystem. It is responsible for scanning the environment for threats. If athreat is perceived, it immediately triggers an emotional and physicalarousal in the limbic system and brain stem. The critical point is thatthis arousal begins to happen before the neocortex has even had achance to register an incoming signal. If the threat is significant, thenthe neocortex is allowed very little chance to intervene. As a result, wemight jump out of the way of an oncoming car or remove our handfrom something hot, say, and only recognize what has happened afterthe event. If, however, the threat is below some (subjective) thresholdlevel of significance, the arousal starts, but is regulated to some extentby the left pre-frontal lobe.

The amygdala, therefore, acts as a trigger to an emergency neuralcircuit that generates an emotional arousal. Its evolutionary functionis obviously to enhance survival prospects, but it is arguable that thetrigger is too sensitive and requires positive intervention to ‘retrain’ it(we shall deal with this in a little more detail in Chapter 22). In themeantime, we need to register the fact that the amygdala is influentialin the formation of crowds.

THE RESPONSE TO A THREAT

Significantly, a crowd is most likely to emerge in the presence of athreat of some kind. A group of people can have a common cause, butwill not be regarded as a crowd, as such, until a threat appears. Acrowd has a certain urgency and immediacy to it, which is not neces-sarily present in a group. Hence, there are two possibilities:

� a number of individuals will feel threatened in some way and thenform a crowd;

� a group already exists that is then sufficiently threatened totransmute into a crowd.

In either case, it is not fanciful to suggest that crowd members’ emer-gency amygdala circuits will be triggered. As a result:

� group survival needs overwhelm individual self-assertiveness needs;


� group beliefs replace personal beliefs;

� the influence of the neocortex is overwhelmed by the influence ofthe brain stem and the limbic system.

Crowds are, accordingly, involved primarily with instincts, biologicaldrives, compulsive behaviour and emotions. Hence, their behaviour isessentially non-rational (and is, in fact, often irrational). To para-phrase Arthur Koestler:20

emotion and intellect, faith and reason, [are] at loggerheads. On the oneside, [is] the pale cast of rational thought, of logic suspended on a thinthread all too easily broken; on the other, [is] the raging fury of passion-ately held irrational beliefs, reflected in the holocausts of past andpresent history.

THE ‘INTELLIGENCE’ OF CROWDS

This certainly helps to explain the popularly held delusion that allmembers of a crowd are unintelligent: it is not that they are unintel-ligent as such – it is that their ability to remain self-aware and thinklogically becomes suppressed in the face of a threat. This is as true ofpairs of individuals confronted with a challenge (such as a couplewhose child-rearing abilities are criticized) as it is of a large group ofpeople facing a physical threat (such as a military unit under fire).21

Whether or not the challenge to a crowd is physical, it almostcertainly involves a threat to its underlying belief system. As a result,the strength of this belief system becomes intensified and severe limi-tations are accordingly imposed on the quality of data that the crowdwill recognize as genuine information. Gregory Bateson definedinformation as ‘differences which make a difference’.22 A crowd mindcan usually only perceive differences that are relatively large and thatoccur over very short periods of time. In other words, a crowd willonly recognize obvious changes. Slow changes can be observed onlyby the lengthy process of continually, and rationally, scanning all thepotentially relevant data. Crowds are incapable of such analysis: theythink in terms of simple images and communicate with slogans. Asthey emerge in times of crisis, they are all too often the main vehiclefor historical ‘progress’.

Two’s a crowd 23

NOTES

1. The terms ‘self-assertive tendency’ and ‘integrative tendency’ were coined byArthur Koestler.

2. Jantsch, Erich (1980) The Self-organizing Universe, Pergamon, Oxford.3. For a very straightforward commentary on this process, see Schumacher, E F

(1977) A Guide for the Perplexed, Jonathan Cape, London.4. This is called ‘homeostasis’, and is the source of the phrase plus ça change,

plus c’est la même chose.5. The hierarchical structuring, and therefore the crowd phenomenon, obviously

also exists in the non-human world. It exists, for example, among birds, fish,ants, bees and lemmings. One of the more unusual examples concerns theability of single-cell amoebae to form themselves into a so-called ‘slimemould’ in order to search for food as a group. See Tyler Bonner, John(December 1959) ‘Differentiation in social amoebae’, Scientific American.

6. Le Bon, Gustave Psychologie des Foules, reprinted (1922) as The Crowd,Macmillan, New York.

7. Ibid.8. Bion, Wilfred (1961) Experiences in Groups, Tavistock Publications, London.

Bion distinguished three types of ‘basic assumption’: 1) basic assumptiondependency, where the group behaves as if its primary task is only to providefor the needs of its members and where the group leader is expected to helpmembers to feel good; 2) basic assumption fight–flight, where there is anenemy that has to be dealt with and where, without concern, group members doas they are told by a leader; and 3) basic assumption pairing, where the groupbelieves that a direct relationship between members and a leader will trigger afuture solution to a current problem. If any, or all, of these assumptions arepresent, it is obviously important that individuals do not question their validity.

9. Dawkins, Richard (1976) The Selfish Gene, Oxford University Press, Oxford.10. Bion, Wilfred, op. cit.11. Bateson, Gregory (1979) Mind and Nature: An essential unity, Wildwood

House, London.12. It is arguable, however, that the specific ability to be aware of one’s own

thinking is hardly ever used. See Bennett, J G (1964) Creative Thinking,Coombe Springs Press, Masham, Yorkshire.

13. See Capra, Fritjof (1982) The Turning Point, Wildwood House, London.14. Peck, M Scott (1983) People of the Lie, Rider, London.15. The frontal lobes require continuous stimulation in order to activate the specif-

ically human qualities of empathy and compassion. Not only is this stimulationapplied irregularly, but the activity of the frontal lobes is all too oftensuppressed by alcohol and drugs. See, for example, Franck, Frederick (1991)To Be Human against All the Odds, Asian Humanities Press, Berkeley,California.

16. Erich Jantsch distinguishes three different types of mental activity, each ofwhich corresponds to one of the different areas of the brain:

‘organismic’ mental activity is the function of the brain stem: ‘reflexive’activity is the province of the limbic system: ‘self-reflexive’ activity is conducted by the neocortex. (See Erich Jantsch, op. cit.)


17. Maclean, Paul D, ‘A triune concept of the brain and behaviour’, in Boag, T, andCampbell, D (eds) (1973) The Hincks Memorial Lectures, University ofToronto Press, Toronto.

18. Wilber, Ken (1983) Up from Eden, Routledge & Kegan Paul, London. See alsoJaynes, Dr Julian (1976) The Origin of Consciousness in the Breakdown of theBicameral Mind, Houghton Mifflin, Boston.

19. Goleman, Daniel (1996) Emotional Intelligence, Bloomsbury, London.20. Koestler, Arthur (1978) Janus: A summing up, Hutchinson, London.21. Scott Peck, op. cit., effectively aligns crowd-type behaviour with the concept

of evil. His examples range from co-dependent couples involved in child abuseto the My Lai massacre. The central point in his analysis is the one made byPaul Maclean: compassion for others is effectively eliminated.

22. Bateson, Gregory, op. cit.

Two’s a crowd 25

26

3

The individual in thecrowd

THE INTEGRATIVE TENDENCY

We can now look at the crowd phenomenon from the individual’spoint of view because it follows that, if a crowd is to form, individualmembers must, in some sense, ‘want’ to join and be willing toconform to behavioural standards imposed by others. No group – andespecially not a crowd – can exist unless each group member alterstheir behaviour to conform to the will of the majority. In other words,there must be a basis for the integrative tendency, or the tendency todevote oneself to a group.

Theoretically, we can observe three aspects to this phenomenon,1namely:

� identification with the crowd;

� acceptance of the crowd’s belief system;

� submission to the authority of a leader.

IDENTIFICATION

There is, indeed, a well-researched predisposition for individuals toreject individuals in other groups, to accept the judgement of themajority in the group to which they belong, and to accept the

The individual in the crowd 27

instructions of a person (or persons) representing authority. In a seriesof experiments by Henri Tajfel2 at Bristol University, it was shownthat groups of schoolboys aged 14 to 15 could have their behaviouraltered merely by telling them that they belonged to a particular group– even an unknown group. Specifically, the schoolboys would auto-matically associate themselves with other members of the samegroup, would provide active support for that group and would takeevery opportunity to disadvantage members of other groups. Thesephenomena occurred despite the fact that no indication was given tothe schoolboys about the purpose or qualities of the groups.

BELIEFS

It seems that uncritical acceptance of group assumptions/beliefs hasits foundation in the subconscious need to belong, or not to be seen asdifferent. For example, experiments conducted in 1956 by SolomonAsch at Harvard3 showed that, when matching the length of a line withone of three other lines, subjects could have their performance meas-urably altered by group pressures. When asked to match the length ofa line with one of three other lines in isolation from others, partici-pants made a mistake less than 1 per cent of the time. However, whenplaced in a group that had been instructed beforehand to claim thatmismatched lines were actually the same, 75 per cent of participantsagreed with the majority. This was true even when the actualdifference between the lines was very significant. Worryingly, Aschalso found that although some participants lacked the nerve todisagree with the majority, some maintained that they had actuallyseen the mismatched lines as being equal and some doubted their ownperceptions.

SELF-AWARENESS AND CONFORMITYENFORCEMENT

These general conclusions have been confirmed by researchconducted at the University of Illinois by Ed Diener. He found that themost important factor in group behaviour was the suppression of self-awareness and, therefore, of self-regulation. In one particular piece ofresearch, Diener compared behaviour under three different laboratorytest conditions:4


� where individuals were self-aware and isolated from groupinfluences;

� where individuals were non-self-aware, but were still isolated fromgroup influences; and

� where individuals were both non-self-aware and involved in agroup environment.

It was found that individuals in the third situation generally actedspontaneously, had little sense of personal identity and related closelyto other group members. Indeed, the experience was found to be veryenjoyable. Quite clearly, conformity enforcement within the groupgoes hand in hand with reduced self-awareness by individuals.

THE CROWD LEADER

Identification with other members of the crowd and an acceptance ofthe crowd’s belief system is stimulated by each member’s willingnessto obey a crowd leader. The importance of a leader for crowddynamics was Sigmund Freud’s major contribution to the debate,5 andis based on his idea of a ‘parent substitute’.6 Specifically, Freud arguedthat a large group (or crowd) would follow its leader because thatleader personified certain ideals and objectives for the group. Groupmembers would essentially disable themselves by projecting their owncapacities for thought, decision making and responsibility taking on tothe leader.

The crowd leader acts as the main interpreter of new information fromthe environment, determines the appropriate tactical response and directsstrategy. Leaders may be dictatorial or democratic, they may beconstructive or destructive, but they will always command the attentionof each of the crowd members. However, leadership may manifest in anumber of different guises. On the one hand, it may be obvious, in thesense that it is vested in a particular individual or in a subgroup of people(such as a committee or board of directors). On the other hand, it may becovert. It may, for example, be vested in the democratic decision of thegroup itself or in the shared system of beliefs held by members of thecrowd, and the code of conduct it engenders. Covert leadership, however,almost certainly requires some form of ‘totem’ on which the crowd canfocus its attention. Throughout history, crowds have been responsive tonational flags, figureheads, icons and statues.

THE FINDINGS OF STANLEY MILGRAM

The classic experiment to discover the limit to which people would beobedient to authority was conducted by Dr Stanley Milgram of YaleUniversity.7 In the experiment, the subjects were ordered to inflictpain on an innocent victim in the interests of an important cause.Authority (or the leader and representative of a group belief system)was represented by a scientist in a white coat who would continuallyurge the subject to proceed with administering electric shocks to athird person. In fact, there was no electric shock involved at all: thesubject did not know it, but the third person (the victim) merelybehaved as if there had been one.

At all stages throughout the experiment, the subject was madeaware of the effect of their actions, both by a dial on the electric shockmachine (which indicated the voltage being administered and thedegree of danger involved therewith), and by the screams and protestsof the victim strapped into a chair. Milgram found that over 60 percent of the subjects were prepared to obey instructions to administerthe highest and most lethal dose of electricity, even after the victimhad given up screaming and was, to all intents and purposes,comatose.

ALTRUISM AND CONFLICTIn the previous chapter, it was noted that the structure of the humanbrain lends itself to the emergence of non-rational responses toperceived threats of any kind. In particular, the brain stem and thelimbic system are subject to a significant stimulus. Significantly, thearousal of these two subsystems at the expense of the neocortex candegrade human behaviour to the extent that some degree of animal-like ‘herd’ behaviour materializes. This is the phenomenon of thecrowd.

There are, obviously, different degrees of intensity of crowdbehaviour, but its hallmark is non-rational and emotional behaviour,common to a number of individuals acting together. It follows, too,that once a crowd has been catalysed, the left pre-frontal lobe of theneocortex is even less likely to operate effectively. As a result, theamygdala of each crowd member is given an even freer rein.

The results – as researchers such as Diener and Milgram haveverified – can be frightening. A crowd led by a strong leader can be atruly potent force. First, people within a crowd develop a sense of



altruism towards other crowd members that is very strong.(Sometimes, indeed, it is so strong that, as Emile Durkheim found, itcan result in suicide.8) Second, the crowd can achieve objectivesusing methods that more self-aware individuals would regard asbeing totally unacceptable. It is not surprising, therefore, that peopleare more likely to be involved in states of conflict as group membersthan as individuals. Third, and as a corollary, it follows that conflict(or stress) is a perfect catalyst for the formation of a crowd. If, forsome reason, an imbalance develops between two groups, each groupmember will have common cause with other members of the samegroup in protecting the autonomy of that group. The paramount needof each group may then release the aggressiveness in, or relaxconstraints on, each individual.

SPLITTING AND PROJECTION

There is a broad question about what encourages, or allows, indi-viduals actively to participate in aggressive, destructive crowdbehaviour when they might not otherwise do so. One factor is theneed to belong. For many, if not all, crowd participants, the sense ofbelonging totally to a higher psychological construct is a transcen-dental experience: ego boundaries collapse and are invaded by groupbeliefs, personal responsibility is removed and anxiety is eliminated.9A second factor is the way that the human psyche deals with poten-tially conflicting emotions. Here, an individual disowns the unac-ceptable side of his or her psyche and projects it out on to others.10

The result is a toxic combination of infallibility on the one hand andhatred and fear of an ‘enemy’ on the other, and it is all too easy to killand maim ‘subhuman others’.11 Unfortunately, the fact that theseprocesses occur in the subconscious mind makes it all too easy todeny that they exist in the first place.

CONCLUSIONThe conclusions of this chapter are simple, but profound. Member-ship of a group in general, or of a crowd in particular, involves theabrogation, to some degree, of personal responsibility – that is, peopleact differently as crowd participants than they do as independent indi-viduals. A crowd as a whole tends to behave in a non-rational,emotional way in pursuit of its objectives and forces its members to

do likewise. The ability of a crowd to organize its members in thisway is particularly pronounced under conditions of conflict when theautonomy of the crowd is in some way threatened.

These observations go some way towards explaining some of theless attractive features of the human condition. They explain, forexample, why armies of otherwise rational and humane men areprepared to go to war; they offer an explanation of why avowedly reli-gious groups have tortured and murdered in pursuit of doctrinalpurity; they explain why trade union members have been willing todestroy the companies that they work for rather than surrender anyunion ‘rights’. The list is endless, and is the more depressing for it.

However, the purpose of this book is other than to bemoan the fateof humanity. Up to now, Part One has presented a body of theory thatexplains, essentially, why group behaviour is a ubiquitous feature ofthe human condition. For the vast majority of people, some form ofgroup pressure provides a major motivating force in all their social,economic and political activities. Such pressure exists in such diversestructures as friendly societies, corporations and religions. It is moreintense in sports teams and is at its most intense under combat condi-tions. As we shall demonstrate, the crowd phenomenon also exists infinancial markets.

So far we have only explained why crowds exist. We have not yetshown how crowds behave. Let us, therefore, take one further steptowards the primary purposes of this book by analysing the dynamicsof a crowd system.

NOTES

1. Koestler, Arthur (1978) Janus: A summing up, Hutchinson, London.2. Quoted in Arthur Koestler, op. cit.3. Asch, Solomon (1956) ‘Studies of independence and conformity’,

Psychological Monographs, 70 (9).4. Diener, Ed, ‘Deindividuation: the absence of self-awareness and self-regu-

lation in group members’, in Paulus, P B (ed.) (1980) The Psychology ofGroup Influence, Erlbaum, Hillside, New Jersey.

5. Freud, Sigmund (1959) Group Psychology and the Analysis of the Ego,Norton, New York.

6. Freud and others have argued that the willingness of individuals to respond toa leader is based on the experiences of early childhood. Under normal circum-stances, each baby not only learns about its individuality, but also gainssecurity from transactions with its parents. Hence, a need for both a parentsubstitute and the company of others is learnt and carried through into adultlife.


7. Milgram, Stanley (1975) Obedience to Authority, Harper & Row, New York.8. Durkheim, Emile (1970) Suicide: A study in sociology, Routledge & Kegan

Paul, London. Durkheim differentiated between altruistic suicide, where indi-viduals die for the group, and anomic suicide, where they die because of sepa-ration from the group.

9. See Neumann, Erich (1990) Depth Psychology and a New Ethic, Shambhala,Boston, Massachusetts. See also Chapter 6, note 8.

10. See Klein, Melanie (1946) ‘Notes on some schizoid mechanisms’, in (1997)Envy and Gratitude, Vintage, London.

11. The psychology of the suicide bomber, for example, involves splitting andprojection. However, it also involves a combination of Durkheim’s anomicsuicide and altruistic suicide (that is, the bomber is simultaneously cut off fromthe wider community and committed to protecting the integrity of a specificgroup or idea). Logically, the only solution to the problem is directly to addressthe initial sense of alienation.


33

4

The systems approach tocrowd behaviour

INTRODUCTION

The insights of analysts such as Le Bon, Freud and Koestler take us along way towards an understanding of the basic crowd phenomenon.However, this does not complete the process. There is one last step totake because, in order to demonstrate the importance of the crowdwithin the context of financial markets, it is necessary to understandhow the crowd generates ‘fluctuations’.

It is possible to do this thanks to the theoretical developments thathave taken place in the analysis of living organisms during the last 20years. Crucially, we can analyse the dynamics of crowd behaviourwithin the framework of the philosophical approach known as‘systems theory’. Such an approach – which stems from thepioneering work conducted by analysts such as Ludwig vonBertalanffy,1 Ervin Laszlo,2 and Erich Jantsch3 – proceeds to a newunderstanding of nature by focusing on processes rather than on struc-tures as such. Hence, each living organism is seen as being a self-organizing system, with a hierarchical structure. Each such structureis responsive to disequilibrium, open to the environment for theexchange of information and energy, and able to process informationand energy. These characteristics broadly mirror Gregory Bateson’scriteria for the existence of ‘mind’.

In Chapter 2, it was established that – theoretically – a crowd couldbe regarded as having a ‘mind’ of its own. Furthermore, a crowd is


created from living organisms. Logically, therefore, a crowd can beseen as a natural phenomenon. This, in turn, implies that a crowd islikely to be much like all other natural phenomena, in the sense ofbeing a dynamic, fluctuating system while it is in existence. Crowdbehaviour is therefore susceptible to analysis using systems theory.

NON-EQUILIBRIUM CONDITIONS

The basic catalyst for the formation of a crowd is a condition of stress,conflict or competition being experienced by more than one personsimultaneously. The catalyst is a condition of ‘non-equilibrium’ thatproduces a threat of some sort. This non-equilibrium condition thenprovides the purpose for the creation of a crowd – namely, the alterationof the disequilibrium in order to benefit the crowd’s membership. Thetargeted change may merely be the attraction of additional resourcesaway from the environment (as in the case of a profit-making corpo-ration) or it may be a more fundamental attempt to impose creativeevolution on the environment (as in the case of social revolution). Acrowd will continue to respond to the implicit non-equilibrium untileither the crowd’s purpose is achieved or (in extreme cases) the envi-ronment itself dampens down the crowd. Obviously, when a crowd’spurpose is fulfilled, equilibrium exists and the crowd members will go ontheir way.

OPENNESS TO THE ENVIRONMENT: THEEXCHANGE OF ENERGY

In order to achieve its objectives, however, a crowd must be ‘open’ to theenvironment for the exchange of both energy and information. Energy iscontinually being used up by crowd members. Consequently, the loss ofuseful energy (known as the build-up of entropy), as individuals leave forwhatever reason, has to be surmounted by continuously attracting newmembers to ‘the cause’. The rate of addition may be greater than, equalto, or less than the rate of exhaustion or departure. Obviously, thecrowd’s ability to survive cannot continue if the rate of arrival of newmembers is less than the rate of departure. This is one of the reasons‘spontaneous’ street riots rarely persist for any extended length of time,even in the absence of large-scale law enforcement.

The systems approach to crowd behaviour 35

OPENNESS TO THE ENVIRONMENT:THE EXCHANGE OF INFORMATION

This exchange of energy is complemented by an exchange of infor-mation.4 The crowd’s need to change, or manipulate, its environmentfor its own ends involves a continuous process of information transferand analysis. The crowd needs to know how its achievements relate toits objectives and how the environment is responding so that counter-vailing adjustments can be made if necessary. This process thereforeinvolves two levels of analysis:

� the nature of the mechanism for transmitting information;

� the inter-relationship between a crowd and its environment.

THE MECHANISM FORTRANSMITTING INFORMATION

The nature of the information transmission mechanism is veryimportant. Often the receipt of information from the environment, andtransmission of this information, is the prerogative of whoever (orwhatever) constitutes the crowd ‘leader’. Essentially, however, anyindividual within a crowd can receive information on behalf of thewhole crowd and can transmit it to the other members of the crowd. AsGustave Le Bon originally argued,5 crowd members are in a constantstate of expectant attention and are therefore vulnerable to suggestion.Moods, feelings and ideas in such an environment are, of course, verycontagious and they spread rapidly. Modern, high-tech communica-tions now ensure that the same effects can be quickly achieved even ifthe crowd is not assembled in one place.

However, the processes of contagion are not necessarily simplistic.Each person not only receives information and transmits it to others:they can also alter it before transmission. The alteration need not onlyapply to the quality of the information, it may also apply to theemotions that accompany the transmission of the information. Hence,even where the information is ‘hard’ data that cannot be directlyaltered, it will be interpreted by individuals, and will almost certainlybe retransmitted in the context of an emotion (anger, greed, sadness,indifference, optimism, pessimism and so on). The end result is thatcrowd members all experience the same beliefs and emotions aboutthe information that is being transmitted.


FEEDBACK LOOPS AND THETRANSFORMATION OF INFORMATION

The idea that a crowd can process, and respond to, information in thisway can usefully be described in terms of what are called ‘feedbackloops’. Feedback loops transform information. The process of trans-forming information means that there is a difference between the inputof information at any stage in the loop and the output of informationfrom the same stage (see Figure 4.1).

The resultant, changed output then becomes an input of informationfor the next stage. Thus, part A may affect part B, part B may affectpart C, and C may feed back the (transformed) information to A. Eachinput of information therefore triggers a process that generates newinformation.6 Interestingly, there are two possibilities here (see Figure4.2):

Positive feedback Negative feedback

Systemresponse

Systemresponse

Time Time

Figure 4.1 Information transformation

Figure 4.2 Feedback loops and adjustment paths


� information is given a positive gain at each stage, thereby leadingto a ‘runaway’ system;7

� information is given a negative influence (usually by some form of‘governor’) at some stage in the chain, thereby leading to an ‘oscil-lating’ system.

OSCILLATING SYSTEMS

Although the presence of a negative feedback loop always createsoscillations, it does not necessarily generate stable ones. In fact, itmay generate one of three types of oscillation – namely, damped,stable or unstable. These three possibilities are shown in Figure 4.3.

THE ROLE OF THE CROWD LEADER

Within any group, the role of the ‘governor’ is essentially performedby the group leadership. The leadership ensures (or tries to ensure)that group behaviour is either magnified by means of positivefeedback or damped down by negative feedback. Where the lead-ership acts to damp down the oscillations of a group, that group

Systemresponse

Systemresponse

Systemresponse

Figure 4.3 Oscillations found in negative feedback loops

operates harmoniously with its environment. Such behaviour istypical of bureaucratic organizations. Obviously, however, there areoccasions when the leadership attempts to magnify the oscillations ofthe group. This is, perhaps, the defining characteristic of a ‘crowd’.The leadership will have a vested interest in magnifying the frictioncreated by members under its control simply because the intention isto change the environment in a fundamental way.

Under circumstances of destabilizing oscillations, the environment(that is, society) itself will attempt to protect its own autonomy bysuppressing the ‘runaway’ behaviour of the crowd: this is the basis ofpolice riot control. However, at an extreme, a mob may enforceevolution via revolution.

THE INTERRELATIONSHIP BETWEEN ACROWD AND ITS ENVIRONMENT

In this sense, change and ‘progress’ come about as a result of adynamic interplay, both between individuals and the group to whichthey belong, and between a group and its environment. Any conflictcreates compensating adjustments. First, therefore, it is true to say thatthe progressive development of any overall system is determinedpartly by the creative components within it. For example, culturaldevelopment in any society is the result of the interplay of creativefree thinking and conservative traditions. The larger whole willinvariably seek to damp down the smaller parts, but ultimately thesmaller parts must have an impact on the whole if any form ofprogress is to be achieved. Recent cases in point are the Women’sMovement and ecologically oriented groups such as Greenpeace.

Second, independent fluctuations in the environment itself can leadto a number of different responses by a crowd. Initially, a change inthe environment imposes stress on the group or crowd. The group istherefore unable to respond properly and it seeks to stabilize itself bychanging its metabolism.8 Hence, for example, an establishedcompany faced with falling sales will respond by accepting lowerprofits, an increase in bank overdrafts and a rise in stock levels.

Subsequently, of course, the group will ‘acclimatize’ itself toenduring changes in the environment by adjusting its operating abilitywithout actually changing its structure.9 Hence, a company faced withfalling sales would eventually need to take positive action in terms ofa cutback in its operations if it was to survive.



Ultimately, the group may adapt to permanent changes in the envi-ronment as a result of ‘revolutionary’ changes. The group moves to anew, but stable, operating structure; in particular, the objectives of thegroup are altered.10 Inevitably, therefore, a company that either antici-pated, or was faced with, a permanent decline in demand for itsproducts would have to diversify its operations if it were to stop itselffrom going out of business. The importance of timing in this process iscovered in a little more detail in Chapter 9.

It is interesting to note the relationship between chance andnecessity in this view of evolution. According to the old Darwinianview, chance mutation is followed by the survival of the fittest.However, from the newer systems viewpoint, chance and necessityare complementary principles. The initial perturbation that creates theinstability in the system may appear to be random from the system’spoint of view,11 but the need to survive forces the system to adapt. Thesystem, in a sense, makes its own decisions about the nature of its newstructure. Evolution is thus the unfolding of order and complexity inthe process of learning.

CONCLUSION

To summarize the analysis of this chapter, we may argue as follows:any human grouping – specifically, a crowd – is part of the hierar-chical structure of nature; and each crowd can be defined in terms ofits processes rather than its physical characteristics. Theoretically,these processes are triggered by differences (or non-equilibriumconditions), utilize feedback loops and therefore involve oscillations.The processes are continually energized by the crowd’s access to theenvironment and are sufficiently complicated to facilitate somedegree of ‘self-knowledge’. Energy and self-knowledge, in turn,contribute to the crowd’s ability to be self-organizing. Self-organi-zation involves the control of the crowd’s members, and this enablesthe crowd to maintain its autonomy as well as to learn, adapt andevolve. In somewhat technical language, we have described thebehavioural characteristics of Le Bon’s ‘crowd’.

NOTES

1. von Bertalanffy, Ludwig (1968) General Systems Theory: Foundation, devel-opment, applications, Braziller, New York.

2. Laszlo, Ervin (1972) Introduction to Systems Philosophy: Towards a newparadigm of contemporary thought, Gordon and Breach, New York.

3. Jantsch, Erich (1980) The Self-organizing Universe, Pergamon, Oxford.4. The environment ‘recognizes’ a crowd by responding both to the difference

made by that crowd’s initial presence and to the difference made by its subse-quent activities; while the crowd itself ‘recognizes’ its own existence byresponding to the resulting differences in the environment.

5. Le Bon, Gustave, Psychologie des Foules, reprinted (1922) as The Crowd,Macmillan, New York.

6. Such systems are called ‘autocatalytic’.7. The growth may be either exponential or hyperbolic. Exponential growth

results in increases that are proportional to the amount present and in which thedoubling time remains constant. Hyperbolic growth results in increases that arethe square of the amount present and in which the doubling time is halved withevery doubling.

8. Metabolism means those chemical changes occurring in living organisms thatare essential to sustain life.

9. Such ‘somatic’ changes, however, are not necessarily permanent because theyare reversible. Consequently, the changes generally amount to an internal-ization of stress.

10. Fluctuations in the components, caused by environmental changes, are re-inforced by feedback loops. The system is then driven into a new structure by‘order through fluctuation’. Jantsch, Erich, op. cit.

11. The perturbation is usually part of a larger cyclical fluctuation. However, itappears to the lower-level structures as a shock (see Chapter 5).


41

5

Cycles in the crowd

INTRODUCTION

The important feature of any self-organizing system, whether it be acrowd or a living organism, is that it oscillates during the transfer ofenergy and information. Indeed, the presence of continuous fluctua-tions can be taken as prima-facie evidence of the presence of ‘mental’activity (as defined by Gregory Bateson). There are, however,different types of oscillation that need to be distinguished, and weshall look at each of these in turn.

THE LIFE CYCLE

The first point to make is that every crowd has both a beginning and anend. There will be a time before the crowd existed and there will be atime after it has disappeared. In other words, each crowd has a lifecycle. Life cycles essentially consist of three phases (see Figure 5.1):

Integrativetendency

Figure 5.1 The life cycle of a crowd


� growth;

� maturity;

� decline.

Birth and growth

Crowds initially come into being as a result of both a change thatcreates a change (that is, an item of information affects a group ofpeople) and an ability to respond purposefully to that change.1 Once acrowd with common values and a common objective has been estab-lished, it will then respond positively to the input of new items ofinformation from the environment. In particular, it will draw adynamic response from the most creative elements of the crowd, espe-cially from the crowd leadership.2 During the growth stage of its lifecycle, a crowd is completely able to maintain its integrity, even in theface of a hostile environment.

Maturity

At maturity, however, the crowd becomes self-orientated andtherefore inflexible. It is able to feel comfortable, basking in the glowof its achievements, and will therefore seek to control the creativity ofits smaller parts rather than the other way round. Consequently, low-level fluctuations are quickly suppressed.

Decline and death

Eventually, however, this rigidity will mean that the crowd is unableto adjust any further to changes in the environment. The expectationsof the crowd are unchanging, and increasingly become divergent fromactual events. There then begins a decline that is marked by internaldiscord, uncertain leadership and hostility towards that leadership.Ultimately comes the shock that causes the crowd members to partfrom each other entirely.

The life cycle is thus complete, and individuals are released toparticipate in other groups, or crowds. Hence generation and degener-ation are succeeded by regeneration in a continuous process.

Information and the life cycle

It is apparent that the receipt of an item of information by a crowd willhave a different effect depending on the stage of its life cycle. Bad

weather, for example, may stimulate the resolve of pickets in the earlystages of a dispute or it may encourage them to pack up and go home ifthe cause already appears to be lost. Invasion of one country by anothermay stimulate heroic resistance or passive acceptance. Innovations maybe enthusiastically embraced (as in the case of microchips) or vehe-mently opposed (as in the case of the Luddite riots).

CO-EVOLUTION

We have seen that natural dynamic systems process information byusing feedback loops. The essential feature of life (as we know it) isthat it uses negative feedback loops and therefore generatescontinuous fluctuations. In essence, negative feedback loops aredesigned to allow any self-organizing system (including a crowd) tocope with changes in its environment so that it can survive. Indeed,there are two options open to it:

� it can seek to correct the original change in the environment directly;

� it can change itself to meet the change.

In either case, there is a mutual development in the system and theenvironment.

This mutual development involves an interesting and importantconcept – namely that of ‘co-evolution’. The early work on the subjectwas conducted by Vito Volterra in 1926, and was then extended byAlfred Lotka3 in 1956. Their basic theory centred on the existence ofa complementary development between predators and their prey:predators respond to changes in both the quantity and quality of theirprey, but the prey species constantly introduce countervailingmeasures to outwit the predators’ developments. Hence, neither sidewins, and both survive.

The theory was subsequently developed into a more completetheory of co-evolution by the American biologists Paul Ehrlich andPeter Raven4 in 1965. They argued that there is a mutual complemen-tarity between different levels of any hierarchical structure: lower-level changes could not occur without higher-level changes, andhigher-level changes could not develop without lower-level changes.

In 1979, James Lovelock5 sought to extend the basic analysis to theidea that the whole of the Earth’s biosphere has developed into itscurrent form by virtue of co-evolution between living organisms andtheir environment. For example, life could not have been created by

Cycles in the crowd 43


the so-called Big Bang if the necessary fundamental elements had notalready been in existence, yet neither could it have been created out ofthe basic elements if the Big Bang had not occurred. Now, of course,oxygen-breathing/carbon dioxide-creating animals could not livewithout oxygen-creating/carbon dioxide-using plants. The eternal‘chicken and egg’ paradox is resolved by the concept of comple-mentary and simultaneous development.

LIMIT CYCLES

The theory of co-evolution relies on the fact that the relevant feedbackloops generate stable fluctuations between a particular system and ahigher-order system. Actually, and more precisely, this relationship existsbetween a particular system and a niche in the higher-order system.Mathematicians have a name for these stable fluctuations – ‘limitcycles’.6

The term ‘limit’ relates to the fact that the system oscillates withinspecific parameters, and the term ‘cycle’ refers to the fact that theoscillations continually return to the same point of departure.7 If thelimit cycle is stable, the oscillations will spiral on to the solution pathfrom a wide range of initial states. If, on the other hand, the cycle isunstable, a disturbance will cause the oscillations to spiral away fromthat solution path. It follows that co-evolving systems utilize stablelimit cycles (see Figure 5.2).

Figure 5.2 Idealized stable limit cycle


Each limit cycle expresses the relationship between two specificvariables: as one variable moves, so does the other. In Figures 5.2 and5.3, the variables are represented as x and y. However, within thecontext of our current analysis, we could assume that the y axis repre-sents an index of crowd behaviour and that the x axis represents anindex of environmental change. As presented, the limit cycle does notactually take a circular shape on the two-dimensional x–y surface(known as the ‘phase plane’); nor does it need to do so. It may be ovalor elliptical, as well as circular. The important point is that the oscilla-tions occur between predetermined upper and lower limits, and theselimits determine the amplitude of the cycle.

LIMIT CYCLES THROUGH TIME

A moment’s reflection will reveal that the two-dimensional limit cyclerelating x and y (or the environment and the crowd) actually unfoldsover time. In other words, the cycle effectively operates in three dimen-sions. Let us assume that the limit cycle unfolds at a constant rate overtime and so include ‘time’as the third dimension in a diagram. Then thesolution path takes the form of a cylinder (see Figure 5.3).

Figure 5.3 Three-dimensional limit cycle


Furthermore, if we now change the perspective of the diagram, andcollapse one of the x–y dimensions (say x, which corresponds to theenvironmental change), then the resulting solution path yields aregular cyclical pattern in the other variable (y, which corresponds tocrowd behaviour). Successive peaks will occur at regular intervals, aswill successive troughs (see Figure 5.4). The regularity of the intervalsbetween peaks and troughs defines the periodicity of the cycle.However, because a limit cycle can take any number of forms, there isno reason for the shape of the cycle to be balanced around the peaksand troughs: the up phase of the cycle may be longer or shorter thanthe down phase. If we collapse the time dimension, of course, we arereturned to the basic limit cycle.

LIMIT CYCLES IN NATURE

There is, in fact, substantial evidence that limit cycles occurthroughout nature. The human body, for example, literally pulsateswith the rhythms of limit cycle fluctuations: the heart beats regularly;the neural activity of the brain proceeds in bursts of pulses;8 theactivity of breathing is rhythmical. Furthermore, Professor RexHersey of the University of Pennsylvania found that each individualhas a personalized emotional cycle lasting for an average of 35 days.9

In each case, lower-level systems are oscillating regularly inharmony with each other and with the higher-level system (which is

Figure 5.4 Two-dimensional limit cycle


the body itself). In turn, the metabolism of the body is integrated withthe wider environment. There is, for example, a natural metabolic rateof activity and rest that harmonizes with the day/night cycle. Hence,the rate of urine production peaks each day at around the same time,10

and the ability of the blood to coagulate on a wound is lowest in themorning. Furthermore, the rate of body metabolism is generallyhigher during the spring and summer than it is during the autumn andwinter. It appears, in fact, that subtle electromagnetic changes in thegeophysical environment induce electrochemical changes in thehuman body.11

MULTIPLE LIMIT CYCLES

There are three important features of all natural cycles. First, cycles atall levels of the hierarchy will harmonize with one another as theprocesses of nature do not allow discord. Second, each different hier-archical level has a different time dimension attached to it, so thathigher-level cycles take longer to develop than do their lower-levelcounterparts. Third, the trend of each lower-degree cycle is formed bya higher-degree cycle. The situation can be presented as in Figure 5.5.For simplicity, it is assumed that only two levels of limit cycle exist,although in theory any number of levels could be included. Figure 5.5

Figure 5.5 Two-limit cycles in three dimensions


shows a small, short-term cycle integrated with a larger, longer-termcycle in three-dimensional space. As before, the three dimensionscould be an index of crowd behaviour, of environmental change andof time. The short-run fluctuations continue throughout each completelong-run cycle.

MULTIPLE CYCLES IN TWO DIMENSIONS

In order to obtain a clearer picture of this theoretical construct, we canvisualize it in two dimensions: first, by collapsing the ‘time’dimension while keeping the other two dimensions unchanged;second, by collapsing the x (or ‘environment’) dimension whilekeeping the other two dimensions unchanged (see Figure 5.6). Theeffect of higher-level fluctuations on the lower-level fluctuationsshould be perfectly clear.

THE IMPACT OF SHOCKS

Although limit cycles are the main mechanism whereby a self-organizing system copes with fluctuations in its environment, they donot fully represent the adjustment processes that are involved. Bydefinition, items of information become available only in discon-tinuous or discrete time intervals. The adjustment process depends onwhether or not the recipient system is prepared for the information. Ifthe information is expected, then the system can adjust very quickly to

Figure 5.6 Two-limit cycles in two dimensions


suit the changed circumstances. If, however, the information is unfa-miliar to the system when it is received,12 then that informationimpacts as a shock and the system may have to change its dynamicstructure in order to cope.

The concept of shocks as applied to group behaviour is not a newone. As early as 1933, Ragnar Frisch13 sought to show that businesscycles could be simulated by subjecting a linear model to randomshocks. However, the theme being developed in this book presents asubstantially different view of the world from that visualized byFrisch and many subsequent economic theorists. In particular, cyclicalbehaviour is here seen as an intrinsic part of nature, whereas the clas-sical view of economics assumes linear behaviour that wouldconverge on a ‘steady state’ if it were not upset by exogenous changes,such as strikes, crop shortages, changes in government policy,changes in tastes and so forth.

In practice, then, shocks are delivered to systems that are alreadyoscillating in a limit cycle pattern with their niche in the environment.The situation can be viewed as a ‘family’ of limit cycles operating inthree-dimensional space. Each subsystem bears a limit cycle rela-tionship with a higher system; each higher system has a limit cyclerelationship with its local environment; the local environment has alimit cycle relationship with the global environment. Divergencesbetween lower-level cycles and higher-level cycles are rectified byshocks, and the lower-level cycle will then move to meet the upper-level cycle.

THE PROFILE OF SHOCKS

The important feature of the analysis of shocks is the form taken by thesubsequent adjustment process. First, a shock delivered to a lowersubsystem destabilizes the relationship between that subsystem and ahigher-level system. Second, every attempt by the subsystem to re-establish its original relationship will now be met by a countervailingresponse from the higher system. Third (and as a consequence), everyadjustment by the higher system will result in an additional responsefrom the subsystem. Fourth, these fluctuations will continue at leastuntil the limit cycle of the next higher degree is able to reassert control.14

This process of adjustment in three-dimensional space reveals itselfas a spiral (see Figure 5.7). Such a spiral could, for example, materi-alize after a crowd had initially been ‘born’ and was moving towards astable relationship with its environment. Alternatively, it could


represent the adjustment processes within the crowd as it responded toan externally imposed shock.

SHOCKS IN TWO DIMENSIONS

If we now review the solution path in two dimensions, two familiarpatterns emerge (see Figure 5.8). First, we have a spiral formationbetween x (which could be, for example, an index of environmentalchange) and y (which could be an index of crowd activity). Second,we have an unstable cycle, involving either x or y through time. As weshall see in Part Two of this book, these two patterns are of profoundimportance to our analysis.

SOME INSIGHTS INTO SOCIAL CHANGE

The dynamics of hierarchical crowd systems therefore involve threetypes of oscillation:

� the life cycles of the crowd systems involved;� the limit cycles;� the unstable fluctuations.

Environmental change

Index ofcrowdactivity

Figure 5.7 The spiral adjustment process


During the life cycle of any particular crowd, its own internal cyclesare bound by the requirements of higher-level oscillations, and diver-gences are rectified by shocks. This conclusion leads us to a numberof important insights into social and economic developments. Thefirst is that, because higher-level cycles are often so slow to develop,the social and geophysical environments usually appear static andimmutable to the majority of participants. Individuals do not thereforerecognize that changes have taken place until some form of catas-trophe occurs. Revolutions, large-scale population migrations andwars are invariably the results of complicated changes that have beendeveloping over long periods of time. However, they are often trig-gered by apparently ‘local’ causes or even by relatively insignificantevents. The resulting gap between personal expectations and thereality of the situation invariably adds to the chaos by imposing stressand uncertainty on individuals in unfamiliar circumstances.

The second implication is that some of the events that are usuallyassumed to be random (such as the ‘acts of God’ invoked by insurancecompanies) may, in fact, be predictable. It will always be difficult toforecast the exact timing and location of some events, but this does notmean it is impossible to forecast that they will occur within specifictime frames. There is now substantial evidence to suggest that suchdiverse occurrences as wars,15 earthquakes and climatic changesexhibit cyclical behaviour. Most individuals do not recognize therhythmic nature of the phenomena because they are exposed to theirconsequences irregularly, if at all, in the course of a single lifetime.

Figure 5.8 The adjustment process in two dimensions


However, the concept of cyclical fluctuations in the hierarchy ofnature implicitly confirms that apparently irregular events are moreregular than we care to suppose.

Third, the analysis highlights the fact that short-term and long-termdevelopments are intricately related to one another. Evolution occurs atall levels, and in all degrees, in a largely complementary fashion.Consequently, the difference between short-term cycles and long-termtrends is not one that only involves quantity, but one that also involvesquality. In other words, dynamic structures are continually changing ina largely irreversible fashion between one situation and another. Aseach short-run cycle swings down, the quantitative retracement of theprevious rise is not entirely matched by a reversal in qualitativechanges. Even in the case of a simple heartbeat, for example, there is asubtle ageing process between one beat and another. In the case ofeconomic activity, each quantitative cycle in inventories (or stocklevels) is accompanied by important changes in the quality of theinvestment. ‘All change,’ it is said, ‘is growth; and growth is life!’

NOTES1. This raises an interesting point. A crowd can only form in response to the

receipt of information that gives the crowd a purpose; but the concept ofpurpose presupposes that individuals are in some way pre-programmed torecognize the validity of a specific objective. Each individual has to be able toplace the information in a context that triggers their integrative tendency: theone is impotent without the other. Contexts are essentially learnt either directlyor through genetic transmission. This confirms the need for learnt commonvalues to ensure that societies survive. It also confirms the ‘biological’ natureof crowd behaviour.

2. If necessary, new common values and traditions will emerge that will ensurethe stability of the crowd.

3. Lotka, Alfred J (1956) Elements of Mathematical Biology, Dover Publications,New York.

4. See Jantsch, Erich (1980) The Self-organizing Universe, Pergamon, Oxford.5. Lovelock, James E (1979) Gaia, Oxford University Press, Oxford.6. A limit cycle is defined as the (isolated) periodic oscillation between two vari-

ables, and is represented graphically by an (isolated) closed non-linear path.See, for example, Jordan, D W, and Smith, P (1977) Non-linear OrdinaryDifferential Equations, Oxford University Press, Oxford.

7. The time between successive troughs or successive peaks is known as the peri-odicity of the cycle.

8. Freeman, Walter J (1975) Mass Action in the Nervous System, Academic Press,New York.

9. Hersey, R B (1931) ‘Emotional cycles of man’, Journal of Mental Science.Hersey found that the cycle of each individual, taken separately, varied quitemarkedly from the average: some had cycles of 16 days, others had cycles of as


long as 63 days. The important point, however, is that each individual had apersonalized cycle that persisted over the sampling period.

10. Watson, Lyall (1973) Supernature, Hodder & Stoughton, London.11. There is an expanding body of literature that records the striking correlation

between electromagnetic fluctuations in the solar system, electromagnetic fluc-tuations in the Earth’s biosphere and electrochemical changes in the humanbody. The implication is that we are more influenced by events such as sunspotcycles and lunar cycles than we sometimes care to believe. See, for example,Gaugelin, Michael (1973) The Cosmic Clocks, Granada, London.

12. That is, there is no previous context within which to place the information.13. Frisch, Ragnar, ‘Propagation problems and impulse problems in dynamic

economics’, in (1933) Essays in Honour of Gustav Cassel, George Allen &Unwin, London.

14. During the adjustment process, the upper-level cycle does not therefore accu-rately reflect the trend of the lower cycle.

15. Some of the evidence for the striking phenomenon of cycles in warfare isdiscussed in Dewey, Edward R, with Mandino, Og (1971) Cycles: The myste-rious forces that trigger events, Hawthorn, New York.

54

6

Approaches toforecasting crowd

behaviour

INTRODUCTION

The idea that behaviour in financial markets is essentially a crowdphenomenon is the basis of a comprehensive approach to accurateforecasting in equity, bond and foreign exchange markets. At oneextreme, a self-aware individual is potentially unpredictable exceptwithin the very broadest of guidelines. However, at the other extreme,people, as a group, are predictable. This is the essence of the crowdphenomenon, and it implies that price movements in financial marketsare also intrinsically predictable.

RANDOM OR NON-RANDOM

First, however, it is necessary to demonstrate that price movements infinancial markets do display some form of underlying order – that is,it is necessary to demonstrate that price movements are not alwaysand everywhere just random fluctuations. This is a difficult task,because the issue of randomness or non-randomness in financialmarkets is still the subject of academic debate. But we need to get a

Approaches to forecasting crowd behaviour 55

general sense of the phenomenon in order to assess whether there is acase for assuming that tradable price trends exist and forecastableprice patterns reproduce themselves. At this stage of the analysis,however, all that is necessary is that we look at the subject pragmati-cally, without going into a detailed discussion of statistical theory, inorder to get a sense of the underlying phenomenon.1

PRICE MOVEMENTS IN THE DOW

Non-random behaviour exists in a financial market if the pricemovement over one period of time is related in some predictable wayto the price movement over a previous period of time. However, thereare two elements to this statement: one is the time periods being used(for example, one day or one month) and the other is the time elapsebetween the periods that are being compared (for example, today withyesterday or this month with the same month last year). Here we shallonly be dealing with the influence of contiguous time periods, notwith the time elapse between them, because we need to focus on conti-nuity rather than discontinuity.

The analysis that follows concentrates on the movements betweenthe daily closes in the Dow Jones Industrial Average Index overvarious periods of time. The current time period will be denoted ‘timet’ and will be plotted on the vertical axis of each chart, and the relevantprevious time period will be denoted ‘time (t – 1)’ and will be plottedon the horizontal axis of each chart. The result is that movements inthe Dow are plotted in what can be called ‘t/(t – 1) phase space’.

Figure 6.1 plots each day’s percentage change in the Dow againstthe previous day’s percentage change, for every trading day between 2 January 2005 and 31 October 2005. The result is exactly the sort ofpattern that could be expected if daily movements were random – thatis, they are scattered widely throughout the phase space. It is notpossible to predict from one location where the t/(t – 1) price combi-nation will move to next. So, it does not appear possible to predictwhat might happen tomorrow from what has happened today.

This conclusion does not change if the analysis is extended to coverlonger periods of time. Figure 6.2 goes back further in time, to coverthe period from January 1990 to October 2005 inclusive. In fact, thelonger time period ensures that additional extreme price movementsare captured. So, the more complete the data, the stronger is theapparent evidence for randomness. However, an increase in the


number of observations has caused something quite interesting toemerge: the price movements are tending to concentrate in a specificarea in the t/(t – 1) phase space. In this case, by far and away thelargest number of daily price movements do not exceed 4 per cent.

This conclusion becomes even more obvious if the horizontal andvertical axes of the chart are significantly extended by quadrupling thescale (see Figure 6.3). By widening our perspective in this way, it

-2.2-2

-1.8-1.6-1.4-1.2

-1-0.8-0.6-0.4-0.2

00.20.40.60.8

11.21.41.61.8

22.2

-2.2 -1.8 -1.4 -1.0 -0.6 -0.2 0.2 0.6 1.0 1.4 1.8 2.2

t-1

t

DJ Industrial Average:Jan 2005 to October 2005

(Daily close, t/t-1)Data source: Economagic

1-day percentage changes

Figure 6.1 The Dow, January to October 2005

Figure 6.2 The Dow, January 1990 to October 2005

-8

-6

-4

-2

0

2

4

6

8

-8 -6 -4 -2 0 2 4 6 8t-1

t

DJ Industrial Average:Jan 1990 to October 2005



becomes very obvious that there is some form of natural limitation onday-to-day price movements. These movements tend not to move outbeyond a very specific two-dimensional region in t/(t – 1) phasespace. Nor does this conclusion change much if the data set isexpanded even further. Figure 6.4 shows the period from January1946 to October 2005 inclusive, with the same highs and lows on theaxes as were used in Figure 6.3. This time period therefore includesthe Equity Crash of October 1987. The chart clearly picks out the




-32

-28

-24

-20

-16

-12

-8

-4

0

4

8

12

16

20

24

28

32

-32 -28 -24 -20 -16 -12 -8 -4 0 4 8 12 16 20 24 28 32t-1

t

DJ Industrial Average:Jan 1990 to Oct 2005



-32

-28

-24

-20

-16

-12

-8

-4

0

4

8

12

16

20

24

28

32

-32 -28 -24 -20 -16 -12 -8 -4 0 4 8 12 16 20 24 28 32t-1

t

16/19 Oct 87




19/20 Oct 87 20/21 Oct 87


experience of 16 to 20 October; but it also emphasizes the fact that thechange in prices was quickly pulled back into the central mass. In thiscontext, the 1987 Crash event looks idiosyncratic. Hence, not only arethere natural boundaries to price changes in a particular market butalso, if price movements break those boundaries, they are pulled backwithin them. In a sense, therefore, randomness could be defined interms of price movements that break beyond natural boundaries.

One important inference from this is that randomness has some-thing to do with perspective: the longer the time perspective beingtaken (or, if you like, the broader the context), the less likely it is thatfluctuations will seem random. Anyone who has traded in financialmarkets for any length of time will know when a price movement is‘unusual’, based on his or her experience.

STRANGE ATTRACTORS

The areas in phase space that exert some form of gravitational pull onactivity are now called ‘strange attractors’.2 A strange attractor is apattern in phase space that is traced out by a dynamic system. The centreof the pattern is analogous to the centre of an orbit, and the pattern itselfhas the potential to be mathematically defined. Hence, a strangeattractor is the stable path of behaviour within an apparently turbulentsystem. As James Gleick observes, in the short term any point in phasespace can represent a possible behaviour for the system, but in thelonger term the only possible behaviour is the attractor itself.3

There is a strong sense, therefore, that the boundaried behaviour ofprice movements in the Dow is consistent with the presence of astrange attractor. So let us take the argument a stage further and look atthe price changes over periods that are longer than one day. Figure 6.5shows an example of the five-day percentage changes in the Dow, int/(t – 1) phase space. Here, the five-day percentage change for aparticular day (at time t) is compared with the five-day percentagechange the previous day (time (t – 1)). What is clear is that theformerly circular ‘bubble’ containing price movements begins tospread out along an upward-sloping diagonal line. The 1987 EquityCrash is still obviously there and, again, it stands out as being unusual.But by far and away the largest number of the t/(t – 1) data points arecontained within a price change range of 8 per cent.

This phenomenon then becomes even more pronounced over longertime periods. Figure 6.6 shows 20-day percentage changes in theDow, in t/(t – 1) phase space. Here, although the upper and lowerboundaries for the price changes increase to 16 per cent, the deviationfrom the diagonal line contracts. Again, the 1987 Equity Crash standsout as being an unusual event, but the vast majority of t/(t – 1) rela-tionships stay within a boundaried area. In other words, 20-day price


-32

-28

-24

-20

-16

-12

-8

-4

0

4

8

12

16

20

24

28

32

-32 -28 -24 -20 -16 -12 -8 -4 0 4 8 12 16 20 24 28 32t-1

t




16/19 Oct 1987

19/20 Oct 1987

20/21 Oct 1987



-40

-36

-32

-28

-24

-20

-16

-12

-8

-4

0

4

8

12

16

20

24

28

32

36

40

-40 -32 -24 -16 -8 0 8 16 24 32 40

t-1

t

16/19 Oct 87




13 Nov Oct 87

19/20Oct 87

changes for time t tend towards equivalence with price changes attime (t – 1).

PREDICTABLE PRICE MOVEMENTS

This conclusion is profound. It means that, despite short-term turbu-lence, movements in the Dow tend towards a stable path of changeover longer time periods. In other words, price movements in the Dowdo have a basic element of predictability in them. But there is another– equally important – point that emerges from the analysis. This is thatprice movements are persistently induced to accelerate and deceleratealong their equilibrium path. We can hypothesize that investors chasethe market higher until some form of upper boundary is reached, andthen chase the market down until some form of lower boundary isreached. Not only is this highly suggestive of a deep-running orderedprocess at work, but it is also consistent with the idea that oscillationsin financial markets have a cyclical rhythm. In other words, pricemovements are theoretically predictable.

METHODS OF PREDICTING PRICEMOVEMENTS

However, deciding that financial markets are predictable is not thesame thing as deciding how such prediction may be arrived at. Inessence, there are two main lines of thought. The first is that financialmarket prices ultimately reflect fundamental values in the economy.According to this view, financial prices are determined by actual andexpected developments concerning these values, and forecastingfinancial prices is therefore a matter of forecasting fundamentals.4 Thesecond view is that most of the actual and expected informationconcerning fundamental values is already discounted by the market atany given point in time. Hence, market price movements tend toforecast developments in fundamental values, and forecastingfinancial market prices therefore becomes a matter of deciding whatthe market is already ‘saying’ about itself.


ECONOMIC FORECASTING

Despite the fact that political and social trends often have to be takeninto account, forecasting financial markets on the basis of funda-mentals is essentially the province of economics.

Economic theory argues that the price of any asset is determined bythe interaction of demand and supply – that is, the price will rise ifdemand rises or supply falls; and the price will fall if demand falls orsupply rises. Hence, the main task of an economist is to isolate theforces that will influence the demand for, and supply of, financialassets over time. The main tools are economic theory itself andstatistics – the former to provide the logical framework and the latter toprovide some form of quantification to the analysis and conclusions.5

PROBLEMS WITH ECONOMICFORECASTING

The different theories concerning the specific relationships involvedwill not be discussed here as they are beyond the scope of this book.However, it is worthwhile commenting briefly on the methodologyinvolved.

The analytical procedures used in economics are still steeped inthe traditions of ‘reductionism’ (which we discussed in Chapter 1).These procedures presume that it is possible to understand allaspects of a complicated economic system by breaking it down intoits constituent parts. As a result, economic models tend to be simpleaggregates of mathematically independent equations and ‘closed’systems that are divorced from the wider environment in which theyoperate. However, even if such models are accurate descriptions ofthe area of activity that they seek to represent, there are two imme-diate sources of forecasting error. First, there are no automaticfeedback loops between the model and its environment.6

Consequently, the internal relationships in the model graduallybreak down over time. Second, the assumptions about the envi-ronment – which are made by the forecaster to drive the model – are,of course, subject to human error.

These ‘technical’ problems with forecasting are, of course, clearlyrecognized by economic practitioners and great efforts have beenmade to compensate for them, both by building increasingly bigger(computer) models to embrace ‘the environment’ and by introducing



ever more complicated submodels to deal with apparent idiosyn-cracies. Unfortunately, however, it is not yet clearly recognized thatthe fault can lie not in the incompleteness of the models, but in thestructure of those models.

PROBLEMS WITH THE CONCEPTUALFRAMEWORK

The essential problem is that economic models reflect a particular beliefsystem (or view) about the way that the world operates. They incor-porate a conviction that objects that can be differentiated from oneanother are also dissociated from one another unless a significant rela-tionship can immediately be observed.7 The problem, therefore, is notjust one of inadequate techniques that can be justified on the grounds ofsimplicity; the problem is specifically one of understanding.

Unfortunately, the conviction that boundaries can only be perceivedbecause the delineated objects are independent of one another is soembedded in the ‘Western’ way of thinking that it is difficult to seehow it can be properly superseded; and the results are turning out to bedisastrous:

� the mind becomes separated from the body and its emotions;8

� the body is separated from the wider environment and the latter istherefore seen as a potential threat;

� because of the role of memory,9 the psyche not only uses staticconcepts to describe the world but also sees all relationships asbeing essentially linear and uni-directional.

The human ego therefore faces a dual illusion. On the one hand, its senseof isolation induces fear and a desire to control the environment. On theother hand, its conceptual tools encourage a belief that it can control theenvironment. It is in no way surprising, therefore, that economic models– the prime purpose of which is to assist in controlling the material envi-ronment – intrinsically incorporate the problems that stimulate thedesire to control. Indeed, economic models do not always help withunderstanding reality. They instead only reflect pre-established beliefsabout reality. Structural assumptions are regarded as being the ‘truth’and the map is therefore mistaken for the territory.10

As a result, there are three significant criticisms that can be levelledagainst modern economic theory:11

� it tends to believe (and not just assume) that individuals behavemechanistically and, therefore, in a linear fashion;12

� it does not recognize that individuals are profoundly influenced bythe activities of other people;

� it denies the influence of swings in sentiment.13

As a result, modern economics fails to recognize clearly that thecombination of all economic activities transmutes into a cohesive,oscillating system with its own set of characteristics.

THE RATIONAL EXPECTATIONS HYPOTHESISThe central point at issue here is economics’ belief in the dominanceof rational behaviour in human affairs. Indeed, in the last 20 years orso, this belief has been progressively refined into the hypothesis ofrationally formed expectations.14 This hypothesis is based on threeinterlinked assumptions:

� individuals do not behave irrationally – for example, they will notordinarily put their hand in a fire;

� individuals learn from their mistakes – that is, if they burn them-selves by putting their hand in the fire, then they will not willinglyput their hand in again;

� individuals arrive at their decisions independently of one another –that is, they do not put their hands in the fire simply becausesomeone else does so, or tells them to do so.

According to this analysis, group behaviour is simply the aggregate ofrational behaviour by all individuals.

However, it does not take much insight to realize that the rational expec-tations hypothesis fails to provide a rigorous analysis either of economicbooms and slumps or of speculative excesses in financial markets.15

Rational expectations cannot, by virtue of its central assumption of inde-pendent decision making, explain why people’s expectations converge ona common consensus that is inherently destabilizing. There must be areason for people to do the same thing at the same time. And there must bea reason for people to do strange things at the same time. How, forexample, can events such as the Dutch ‘tulip mania’of 163416 be regardedas the outcome of rational decisions made by independent individuals?There must, in short, be a very good reason for markets to diverge fromperceived fundamentals for sustained periods of time.


BUBBLES AND CRASHES

Part of the answer may be found in the similarities between marketsthat have ‘bubbled’ and then crashed. Figure 6.7, for example, showsthe loci of price action in both the Dow Jones Industrial Average fromJanuary 1921 to July 1934 and the NASDAQ Index between September1995 and September 2002. The time periods involved are obviouslydifferent: the former covers a period of over 13 years, while the lattercovers a period of seven years. However, when the time elapse relatingto the Dow is placed on the lower horizontal axis and the time elapse ofthe NASDAQ is placed on the upper horizontal axis, the pattern ofacceleration into their respective peaks and the patterns of their subse-quent collapses are very similar. The collapses are the Wall Street Crashof 1929–32 and the NASDAQ Crash of 2000–02.

Importantly, this similarity is not an isolated event. Figure 6.8shows the performance of the Nikkei Stock Index between January1970 and April 2004 and then overlays on it the behaviour of theNASDAQ Index between November 1984 and October 2005. Thelong, slow build-up is readily apparent in both indices. Then comesthe clear start of an acceleration that takes both markets into a peak.And then comes the collapse. The patterns are almost identical – atleast in terms of the big moves. It is only after the collapse that themarkets start to produce divergent patterns.


0

50

100

150

200

250

300

350

400Ja

n-21

Jul-2

1

Jan-

22

Jul-2

2

Jan-

23

Jul-2

3

Jan-

24

Jul-2

4

Jan-

25

Jul-2

5

Jan-

26

Jul-2

6

Jan-

27

Jul-2

7

Jan-

28

Jul-2

8

Jan-

29

Jul-2

9

Jan-

30

Jul-3

0

Jan-

31

Jul-3

1

Jan-

32

Jul-3

2

Jan-

33

Jul-3

3

Jan-

34

Jul-3

4

Dow

inde

x

500

1500

2500

3500

4500

5500S

ep-9

5

Dec

-95

Mar

-96

Jun-

96

Sep

-96

Dec

-96

Mar

-97

Jun-

97

Sep

-97

Dec

-97

Mar

-98

Jun-

98

Sep

-98

Dec

-98

Mar

-99

Jun-

99

Sep

-99

Dec

-99

Mar

-00

Jun-

00

Sep

-00

Dec

-00

Mar

-01

Jun-

01

Sep

-01

Dec

-01

Mar

-02

Jun-

02

Sep

-02

NA

SDA

Q in

dex

The Wall Street Crash andthe NASDAQ Collapse

(Weekly closes)Data sources: Economagic & Datastream

Dow Jones Industrials(l.h.scale, lower time axis)

NASDAQ(r.h.scale, upper time axis)

Figure 6.7 Bubble and crash in the Dow and the NASDAQ

NON-LINEAR MATHEMATICSThese similarities have been recognized by Didier Sornette, aprofessor of geophysics at UCLA. In 2003, Sornette demonstratedhow non-linear mathematics can track a stock market bubble andpredict the subsequent crash.17 He arrived at two distinct conclusions:1) the price acceleration into the final peak of a bubble is curvilinear,and the time elapse of oscillations around that accelerating trend getsprogressively shorter; and 2) this phenomenon only works because ofthe impact of ‘cooperative self-organization’. In other words, non-linear mathematics can predict the timing of the peak (because oscilla-tions become so fast that they effectively converge on zero), but suchnon-linear mathematics only work because stock markets are ‘natural’systems, derived from the influence of the crowd.

THE CHALLENGE TO ECONOMIC THEORY

Didier Sornette’s analysis provides a direct challenge to economictheory, not only in terms of providing a mechanism that explains stockmarket excesses, but also right down at the level of basic assumptionsabout personal motivation. And he is not alone. In recent years, thesubject has received detailed attention from a host of psychologists,


0

1000

2000

3000

4000

5000

6000

7000O

ct-8

4

Oct

-85

Oct

-86

Oct

-87

Oct

-88

Oct

-89

Oct

-90

Oct

-91

Oct

-92

Oct

-93

Oct

-94

Oct

-95

Oct

-96

Oct

-97

Oct

-98

Oct

-99

Oct

-00

Oct

-01

Oct

-02

Oct

-03

Oct

-04

Oct

-05

Oct

-06

Oct

-07

Oct

-08

Oct

-09

Oct

-10

NA

SDA

Q In

dex

0

5000

10000

15000

20000

25000

30000

35000

40000Ja

n-70

Jan-

72

Jan-

74

Jan-

76

Jan-

78

Jan-

80

Jan-

82

Jan-

84

Jan-

86

Jan-

88

Jan-

90

Jan-

92

Jan-

94

Jan-

96

Jan-

98

Jan-

00

Jan-

02

Jan-

04

Nik

kei i

ndex

Nikkei(r.h.scale, upper time axis)

NASDAQ(l.h.scale, lower time axis)

Start of acceleration

The Nikkei Crash andthe NASDAQ Collapse

(Weekly closes)Data sources: Economagic & Datastream

Figure 6.8 Bubble and crash in the Nikkei and the NASDAQ

and the conclusions are not in accord with the assumptions (andbeliefs) of theoretical economists. The rich variety of potentialresponses, processed through the dimensions of logic and intuition, isnow well attested. How then can economists assert that people actingindependently of one another will tend to respond, on average, in ascientifically reproducible way to particular stimuli?

Economics’ own (technical) answer is statistical. According to theso-called ‘law of large numbers’, which we mentioned in Chapter 1, alarge number of apparently random and unrelated events produces anoutcome that is predictable. Interestingly, however, the reason the lawof large numbers actually works is usually ignored. It is taken as beinga law of nature and left at that. However, the law hints at somethingboth amazing and important: it suggests that order develops out ofrandom fluctuations. In other words, unpredictable behaviour by indi-viduals converges on predictable behaviour by groups, not because ofthe mechanistic workings of mathematics, but because of thecontrolling influence of the groups themselves.

This implies that economic theory necessarily needs to face up tothe prospect of a revolutionary ‘paradigm shift’ – a shift in thestructure of its beliefs and underlying assumptions about the way theworld works.18 At best, it can be argued that economic theory does notreally understand the nature of economic fluctuations in any mean-ingful way; at worst, that the inability of economic theory to solve theapparently intractable problems of the current era – unemployment,price instability, global pollution and environmental destruction – isactually the consequence of a complete misunderstanding of theforces of nature in general and the human mind in particular.

TECHNICAL ANALYSIS

It should come as no surprise, therefore, to find that economic fore-casts cannot accurately predict either the levels at which market priceswill turn or the time they will do so. Financial markets are fast-moving, continually oscillating, reflections of the processes of trans-formation and change. They cannot be defined by static concepts andlinear relationships. The practical value of economics is thereforelimited to providing background forecasts against which to judge thecurrent trends in markets.

The alternative method of forecasting financial market behaviour isknown as ‘technical analysis’, and it accepts the dynamic structure of


financial markets. It has no need to take account of fundamentalvalues because it is assumed that investors’ expectations concerningthose values (and a lot more information besides) are alreadyreflected in the prices. To paraphrase Oscar Wilde’s view of the cynic,a technical analyst therefore knows the price of everything and thevalue of nothing. This implies that financial markets will, in fact,always be trying to anticipate the future and that, therefore, changesin financial market prices precede changes in fundamental conditions.In many cases, it should be possible to use price behaviour to forecastfundamentals, rather than the other way around.

THE PAST AND PRESENT AS A GUIDE TOTHE FUTURE

This, in itself, would be a sufficient reason for focusing attention onthe behaviour of financial asset prices. However, the claim of tech-nical analysis is that it is actually possible to forecast the futureperformance of a particular market entirely by reference to the actualand historical performance of that market. In other words, no externalfactors need to be included in the analysis.19

For many years, analysts have recognized that certain pricepatterns have a predictive value, that the extent of some price move-ments can be calculated in advance and that there are regular pricecycles operating in some markets. More recently, the use of technicalanalysis has become an increasingly large part of the investmentdecision-making process because it is both simple and profitable. Thebig problem, however, is that very few people understand why theforecasting techniques actually work. This has created tremendousdifficulties of communication within the investment industry. On theone hand there are those who know that technical analysis works, butcannot explain why it does; on the other hand there are those whoknow that fundamental analysis does not work to the required degreeof accuracy, but continue to use it because it is at least explicable.

THE RATIONALE BEHIND TECHNICALANALYSIS

The book up to now should have given a strong indication of the raisond’être of technical analysis. Natural forces encourage people to


indulge in group behaviour. Groups behave as single organisms: theytherefore respond in a predictable way to information shocks, havemetabolic (emotional) cycles and follow a definable path of growthand decay. Unlike any other crowd, however, the behaviour offinancial market crowds is clearly reflected in simple and specific indi-cators. These are the price movements themselves and certainmechanical indices of the underlying activity and energy of the crowd,such as trading volumes.20 Logically, these indicators should reflect theoperation of the appropriate ‘natural’ laws.

We shall therefore now demonstrate that price movements andindices of investor activity actually do behave in accordance with theanalytical framework presented so far. We shall show that all pricemovements are part of a very simple pattern that is the response toinformation shocks, and that prices oscillate rhythmically in responseto the metabolic fluctuations of the crowd. We shall show how thecohesiveness of the crowd ensures that each price movement is math-ematically related to preceding price movements. We shall thereforeshow that it is very easy to differentiate between trends and turningpoints.

NOTES

1. The following method of analysis was suggested by economists WilliamBaumol and Jess Benhabib. See Baumol, William and Benhabib, Jess (1989)‘Chaos: significance, mechanism, and economic applications’, Journal ofEconomic Perspectives, 3.

2. The term was first coined by David Ruelle and Floris Takens in 1971. SeeGleick, James (1988) Chaos: Making a new science, Sphere Books, London.

3. Gleick, James, op. cit.4. Because, on the face of it, new information tends to become available randomly

and investors are assumed to make their decisions independently of one another,current economic analysis tends to accept the hypothesis that stock prices arerandomly determined (the so-called ‘random walk hypothesis’). In fact, stockprice movements can appear to be uncorrelated with one another over particulartime periods. However, the hypothesis is essentially no more than a truism: pricemovements appear to be random because of the prior assumptions of randominformation shocks and random individual behaviour. The hypothesis fails torecognize the ordered nature of the underlying process and take account of actualinvestor behaviour. Furthermore, it has been shown that an unusually long run ofcorrelated price changes occurred during the October 1987 ‘crash’ in the UnitedStates (see Santoni, G J (May/June 1988) ‘The October Crash’, The FederalReserve Bank of St Louis Review). The random walk hypothesis cannot thereforeexplain the crash.

5. See, for example, Christ, Carl F (1966) Econometric Models and Methods,John Wiley, New York.


6. This exacerbates the problems of trying to forecast the effects of shocks.Shocks are continually occurring and, indeed, they are the main catalyst forchange. They can be caused by events such as wars, bad weather, interruptionsto supplies of raw materials, labour disputes and strikes, and bankruptcies.However, ‘closed’ linear economic models cannot be used to forecast theeffects of such shocks with any degree of accuracy. Not only is each shockassociated with a change in the environment within which economic andfinancial activities have to take place, but people’s attitudes to such activitiesare likely to vary. Therefore, it is usually possible to calculate only thedirection of economic changes caused by a particular shock.

7. The tendency of the modern human mind to equate differentiation with disso-ciation is carefully explained in Wilber, Ken (1983) Up from Eden, Routledge& Kegan Paul, London.

8. The ego itself fragments into an acceptable ‘persona’and an unacceptable (andtherefore deeply repressed) ‘shadow’. The distinction between the acceptedpersona and the unaccepted shadow is one originally made by Carl Jung. Theshadow consists of potential behaviour patterns that are suppressed by culturalimperatives. The problem is that the shadow still asserts its power in the formof negative emotions and habitual patterns of denial. A breakdown of themechanism of suppression (for whatever reason) can result in apparentlypsychotic behaviour. It is, for example, arguable that the social disorder inEastern Europe in the 1990s had its roots in suppressed cultural needs. SeeNeumann, Erich (1990) Depth Psychology and a New Ethic, Shambhala,Boston, Massachusetts.

9. See Whyte, L L (1950) The Next Development in Man, Mentor, New York. Thepoint here is that the past is static to us. Once in our memory, it re-presentsitself to us as unchanging and immutable. Thought processes that rely onmemory, therefore, naturally adopt static constructs.

10. Alfred Korzybski used the analogy of a map being equated with the territory toexplain the tendency people have of associating themselves directly with theirmental constructs rather than with the object of their mental constructs.

11. Modern economics theory is broadly referred to as ‘neo-classical’ economics.It is based on two hypotheses: rationally formed expectations (after J F Muth)and the existence of a natural rate of unemployment (after M Friedman). SeeTobin, James, ‘Stabilization policy ten years after’, in (1980) BrookingsPapers on Economic Activity, 1.

12. Linear relationships are usually used even in those mathematical models thatare able to create cyclical fluctuations. A great variety of linear relationshipsallows for oscillations. These include derivatives, integrals, fixed delays andlags of various distributed forms. See Allen, R G D (1967) Macro-economicTheory, Macmillan, London.

13. Economic theory assumes that (uncontrollable) sentiment is not a criticalfeature of economic behaviour. The assumption is, of course, part of a beliefsystem that denies that the human psyche is subject to (unacceptable) passions.It follows, therefore, that economics regards mass mood swings as beingpsychotic and abnormal. It is true that excesses of optimism and pessimismoccur relatively infrequently and so it is certainly consistent with the evidencethat emotional upheavals are unusual. However, manias and panics areactually special examples of a general phenomenon rather than idiosyncrasiesas such.


14. See note 11.15. In the late 1980s, much academic energy was spent trying to integrate the 1987

equity crash with the rational expectations hypothesis. One of the main lines ofthought was to argue that the preceding ‘bubble’ was based on irrelevant infor-mation. No one thought to ask why the majority of market participants – manyof whom spend large amounts of money on research – did not recognize thatthe information was irrelevant in the first place. See Diba, B T, and Grossman,H I (September 1988) ‘The theory of rational bubbles in stock prices’, TheEconomic Journal, 98. The fact is that market participants were receiving veryrelevant information – namely, the behaviour of the market itself. (See alsoChapter 7.)

16. It is almost impossible to visualize a craze for owning rare tulips occurring in acountry that actually produces them en masse, but it happened. Homes andproperties were sold to take advantage of the speculation. See Mackay, Charles(1980) Extraordinary Popular Delusions and the Madness of Crowds,Harmony Books, New York.

17. Sornette, Didier (2003) Why Stock Markets Crash, Princeton University Press,Princeton, New Jersey.

18. See Kuhn, Thomas (1962) The Structure of Scientific Revolutions, Universityof Chicago Press, Chicago. Our understanding of reality usually onlyprogresses as a result of common consensus. Experiments are formulatedwithin the context of this consensus, and data are interpreted according to itsbeliefs and injunctions. There is therefore always a limitation on the flow ofgenuinely new ideas, and existing theories are not easily replaced. Paradigmshifts therefore usually only occur after a crisis of understanding.

19. In practice, some knowledge of fundamental influences is clearly important indetermining the validity of any longer-term forecasts that are produced bytechnical analysis. In Chapter 22, it will be argued that the golden ratioprovides a crucial link and boundary between technical and fundamentalinfluences.

20. Financial markets are therefore an ideal source of information about all crowdbehaviour.


71

Part Two

The dynamics of thebull–bear cycle

72


73

7

The stock market crowd

INTRODUCTIONThe physical structures of any financial market and the institutionalarrangements for settling transactions in that market vary throughoutthe world. Many markets are centred on physical arenas known as‘exchanges’; others are established purely by telephone conversationsbetween willing buyers and sellers. All, however, have common char-acteristics. The first of these, strangely enough, is that it is actuallyvery difficult to identify a physical body of people that constitutes themarket ‘crowd’. The investment community is a far larger, far moreamorphous and far less tangible construct than the group of peoplewho are physically on the floor of an exchange or actually transactingbargains over the telephone at any given time.

THE INDIVIDUAL INVESTORThe basic unit in the investment hierarchy is the individual investor.Essentially, an investor is anyone who has a ‘market exposure’. In thiscontext, the term ‘market exposure’ has two meanings because it mayreflect a view either that market prices are going to rise or that marketprices are going to fall. In the former instance, an investor will be a‘bull’ (or will be ‘long’) of a market, while in the latter instance aninvestor will be a ‘bear’ (or will be ‘short’) of a market. Hence a‘bullish investor’ will either own an equity, bond or commodity orhave a ‘right’ to own an underlying security at some future date.1 On

the other hand, a ‘bearish investor’ will have either a direct holding ofcash (or its equivalent, such as a bank deposit or a building societyaccount) while awaiting an opportune moment to purchase a security(or securities) or will hold an investment that confers the ‘right’ to sellstock at some future date.2

The individual investor may be dealing on their own behalf (as aprivate investor) or be dealing on behalf of an organization (as an insti-tutional investor). The investor’s objective is to maximize total returns,subject to a preferred split between capital and income. The reasons forthe separation of capital and income need not concern us unduly here.Suffice to say that each investor faces a fixed tax regime and a futureset of liabilities (dependent on their age3), and has a natural attitudetowards the acceptance of risk. The tax regime will help to determinewhether or not capital is preferred to income over short periods, and thefuture liabilities will help to determine whether capital is preferred toincome over longer periods. However, the most important consider-ation will be the extent to which an individual is prepared to accept thepossibility of uncertain capital gains as opposed to the probability of acertain income,4 whatever the time period concerned.

THE DEALING STRATEGYThe strategy used by the individual investor to achieve a maximumtotal return is that of buying securities before the market price risesand selling them after the price has done so. There are slight variationson this theme, depending on whether performance is judged againstthe rate of return on short-term money deposits or against a ‘fullyinvested’ position. Private individuals fall into the former categoryand will simply aim to buy at low prices and sell at high ones. Someinstitutional investors employed by pension funds and insurancecompanies, on the other hand, fall into the latter category. These indi-viduals may simply hold up purchases within limited time frames toensure that they are at least buying cheaply. Alternatively, they mayswitch investments from underperforming securities to better-performing securities. For all groups, however, the basic philosophyremains exactly the same: ‘buy cheap and sell dear’.

THE FINANCIAL MARKET CROWDThe next higher level in the investment hierarchy consists of twogroups of investor: those who are bullish and those who are bearish.

74 The dynamics of the bull–bear cycle

As we have seen, bullish investors expect prices to rise, while bearishones expect prices to fall. Obviously, time plays an important part inthese expectations. Investors who trade intra-day may have a timehorizon of minutes or hours, while long-term investors are likely tohave time horizons measured in months. So, in principle, we can alsopostulate a time-based hierarchy of bull and bear groups. Forsimplicity, however, we shall hereafter in this chapter assume thatonly one time horizon exists – a long-term, multi-month, timehorizon. In this way, we can keep potential complications to aminimum without jeopardizing the essential message.

Now, expectations about the future trend in prices necessarily existwithin the context of the most recent trend in the market. On the onehand, there will be a critical mass of participants who have takenadvantage of that most recent trend and will therefore consider them-selves well placed to take advantage of a continuation of the sametrend. These investors will be a ‘successful’ group. On the other hand,however, another group of people will have missed a significant partof the trend and will not be well placed if that trend continues. Thesepeople will constitute the ‘unsuccessful’ group. The unsuccessfulgroup will necessarily consist of two subgroups – namely, those whoare actually holding ‘losing’ investments because they have been posi-tioned against the market trend and those who are ‘disinterested’ in sofar as they currently have no intention of participating in the trend(although they may nevertheless do so at a later date).5

Note, therefore, that it does not matter whether prices have beenrising or falling. The critical issue is whether or not investors havebeen correct in anticipating the trend. It is the relationship betweenexpectations and outcomes that not only determines individualinvestor attitudes, but also determines the psychological environmentwithin which investment is taking place.

We have seen that crowds are not necessarily physical assemblies,but are psychological phenomena. This is especially true of financialmarket crowds. Actual and potential market participants are alllinked to each other via the national and international communi-cation networks. Newspapers, television, telephones and marketreports all supplement the sort of direct contact that habituallyoccurs in daily professional and personal life. This network ofcontacts ensures that price-sensitive information is speedily dissem-inated and assimilated and that interested individuals are, in someway, brought into contact with each other. Understandably, however,the psychological reaction of each individual is going to depend onwhether or not they are being successful.

The stock market crowd 75

THE INFLUENCE OF EMOTIONS

When an individual buys or sells securities, an emotional commitmentis being made. Initially, of course, this commitment is being made to atrading position. The decision to deal may have been arrived atrationally, but the act of dealing creates a financial involvement and anassociated need to ‘get it right’. The investor will have a tradingposition, the market value of which is outside their control. There willbe a feeling of pleasure as the price goes in their favour, but there willbe a feeling of displeasure (dismay, anger, depression, fear) if theprice does the opposite. These feelings of pleasure or displeasure willbe associated with physiological changes in the body – the heartbeatchanges, the rate of respiration alters and the palms sweat.

THE HERD INSTINCT

These feelings are intensified when an individual associates withother people. If the trading position is ‘right’, the personal advantagesin terms of wealth and self-esteem are supplemented by communi-cating with other investors who have similar trading positions.Conversations with others will confirm both the validity of the tradingposition and the decision-making process that preceded it; satisfactionwill be felt when newspaper articles or brokers’ reports providesupportive evidence; members of the same group will engage in acontinuous process of congratulating each other; attention will befocused on the immediate future, where critical analysis is lessnecessary, rather than on the longer term.

More than this, however, members of the ‘successful’ crowd willtend to emphasize the weakness of the arguments of the ‘unsuc-cessful’ crowd, and there will be a continuous propaganda streamaimed at the latter. Those members of the unsuccessful crowd whohave wrong (losing) positions will already feel at a disadvantage andwill be very vulnerable to negative information. Initially, they will feelthe need to associate with others who made the same decision forprotection. They will try to confirm each other’s views as beingcorrect and will emphasize the ultimate errors of the ‘successful’crowd’s case. Indeed, there will be a tendency to ignore the‘successful’ crowd’s arguments altogether. Members of the ‘losing’crowd will commiserate with one another and emphasize the longerterm rather than the immediate future.


Meanwhile members of the ‘disinterested’ subgroup will merelywatch from the sidelines. It would be wrong to apply the term ‘crowd’to this group because there is no emotional commitment to investmentpositions. Nevertheless, this group contains an important source ofpsychological and financial resources that can subsequently beutilized by the ‘successful’ crowd. Ultimately, therefore, this group isalso susceptible to propaganda.

It is apparent from these comments that a commitment to a tradingposition is only the beginning of the story. No matter how rational theoriginal decision to enter into a trading position may be, the very act ofdealing moves an individual into a less rational, crowd-type envi-ronment. Specifically, the investor accepts a specific belief about thefuture trend in prices (that is, that they will go either up or down) andidentifies strongly with those who hold the same belief. They, in sodoing, become committed crowd members.

THE MECHANISM OF PRICE FLUCTUATIONS

The presence of two crowds – the bullish crowd and the bearish crowd –in a financial market, that have diametrically opposite views, ensuresthat a state of conflict exists within the rules of the investment game.While the conflict is balanced over shorter time periods, the marketprice action will tend to take the form of a gentle oscillation. Thefinancial resources being allocated to buying and selling securities willnot generate a long-term imbalance between the two groups. When theconflict becomes imbalanced, however, then one of the two crowdswill start to become successful and the other will become unsuccessful.Consequently, the conflict intensifies and stress levels increase.

As we have already seen, conflict and stress are the main catalystsfor the formation of a crowd mind. In fact, if everybody alwaysthought the same thing at the same time, there would be no ‘market’ assuch and, consequently, no graduated price movements: prices wouldjump up and down randomly, if they changed at all, and no one wouldtrade. However, the influence of differing views as to the future courseof prices and the fact that the conflict between the successful andunsuccessful crowds takes time to resolve ensures that prices trendthrough time.


THE BULL–BEAR LIFE CYCLE IN EMOTIONS

The intensity of this emotional commitment to a particular crowd seemsto vary depending on the phases and duration of a particular bull–bearcycle. We shall be analysing these different phases in more detail in laterchapters. However, the following comments can be made now.

In the early stages of a new price trend, the majority of investorsstill remain committed to the old price trend. The previous trend hasalready finished, but its termination has not yet been generally recog-nized. A minority of investors will have begun to suspect that a pricereversal is at hand and will be altering the balance of their portfoliosaccordingly. However, even for this group, there will be some degreeof uncertainty about the future course of prices: portfolio balanceswill be altered to only a relatively small degree and there will be atendency to reverse profitable trades very quickly rather than take thelonger-term view. At the beginning of a bull trend, the fear of makinglosses still predominates, while at the start of a bear trend greed (or afear of missing further profits) tends to stop people from selling asignificant proportion of their investments. Consequently, in the earlystages of a price trend, the emotional commitment of the (ultimately)successful crowd to that trend is quite weak.

Subsequently, however, there comes a stage when the emotionalcommitment to the developing new trend becomes more intense. This isthe stage when most of the investment community recognize that themovement in prices has become a trend, either upwards or downwards,and emotional conviction replaces any rational doubts that still persist.Investors therefore ‘chase’ the trend and open new trading positions.

This feature of investor behaviour then leads on to the next phase ofthe cycle. Once there has been a general recognition that a bullish orbearish price trend is in full swing, the foundation is laid for a pricereversal. The trend will obviously persist for some time, but the factremains that the more people who commit themselves to believing in aparticular trend and to investing in that trend, the fewer people there arein fact left to perpetuate it. A price reversal therefore inevitably occurswhen the vast majority of investors believe that it will not (yet) happen.

THE OBJECTIVES OF TECHNICAL ANALYSIS

It should be apparent from this (admittedly brief) analysis that theobvious rule for successful investment is to keep a close watch on


what other investors are saying and doing and, then, when a vastmajority are saying and doing the same thing, do the reverse. This isone of the most valuable aspects of using technical analysis: it is arational approach to a non-rational phenomenon and encourages theuser to stand aside from the crowd pressures.

THE INFLUENCE OF PRICE MOVEMENTS ONCROWD PSYCHOLOGY

There are two important points to remember here. The first is thatcrowds have objectives of their own, even if the individual membersdo not specifically recognize them as such. The second is that crowdsrespond very quickly and simply to leadership. It is an importantfeature of financial markets that prices play a crucial part in both ofthese aspects of crowd behaviour. First, it is the prime objective of thesuccessful crowd to continue to move security prices in its ownfavour; second, the leadership role is partly provided by movements inprices themselves. Let us now analyse each of these forces in turn.

THE CONTEST BETWEEN THE TWOCROWDS

The ability of the successful crowd to move security prices in its ownfavour depends primarily on the financial resources that can bemarshalled and committed to the market. Eventually, the conflictbetween the successful and unsuccessful crowds is resolved when thelatter disintegrates and transfers its resources to the successful crowd.

The strategy used to win this conflict is, first, to ensure that everymember of the successful crowd is fully committed to membershipand, second, to bombard the opposition (which includes all those whoare not yet participating in the market at all) with a continuous streamof propaganda. Hence, as a persistent move by prices in a particulardirection develops, the integrative tendency of the successful crowdwill ensure that each member’s trading position is as large as possible.Rational behaviour by the individual becomes more difficult and non-rational (emotional) behaviour becomes increasingly easy. Tradingpositions are opened on the basis that the successful crowd’s beliefsystem (be it rising or falling prices) will continue to be correct and


critical judgement wanes accordingly. Personal interest ultimatelybecomes subordinated to the group interest and overtrading begins tooccur.

THE INFLUENCE OF PRICES ON BEHAVIOUR

As more and more resources are committed to the successful crowd’spoint of view, prices will continue to respond favourably. This, in turn,justifies the successful crowd’s existence and excites members of thatcrowd further. Simultaneously, of course, the vested interests involvedin keeping prices moving in one direction will be increased.Consequently, brokers, market makers and investors alike willprovide a continuous stream of favourable comments to the media,and the latter will promote the cause to the widest possible audience.

Meanwhile, individuals in the unsuccessful crowd will already bevulnerable to coercion and will tend to desert to the successful crowd.As this process develops, market prices will continue to move in thedirection required by the successful crowd. These changes in pricesare a clear signal6 to members of the unsuccessful crowd that theirarguments are ineffectual. No matter how powerful these argumentsmight appear from a longer-term viewpoint, the short-term messagecannot be ignored. The beliefs of the unsuccessful crowd are shown tobe incorrect.

Eventually, the stress created by the combined forces of adverseprice movements, shrinking numbers and unfavourable propagandabecomes too much and the trickle of deserters from the unsuccessfulcrowd becomes a flood. Prices will suddenly rise or fall very sharplyas this change occurs: it is an emotional period, usually accompaniedby high-volume turnover in the marketplace.7 Furthermore, thoseinvestors who join the successful crowd at this stage are ‘converts’and, as such, tend to be the more committed. They are therefore theleast likely to change their minds quickly and so will provide suffi-cient finance to keep the trend intact.

THE LIMIT CYCLE BETWEEN PRICES ANDBEHAVIOUR

From this analysis, it is apparent that price movements are not just apassive response to market forces: there is a feedback effect that ensures


that market forces are themselves responsive to price movements. Inother words, there is a limit cycle relationship between prices andinvestor behaviour. This observation is important and is in direct contrastto the assumptions of classical economics. (Indeed, within the classicalframework, prices cannot be determined if the demand and supply func-tions are allowed to shift in response to a change in prices.) Crowdbehaviour ensures that a price movement triggers an emotional responseamong members of opposing crowds and thereby helps to ensure that themost recent movement in prices is continued into the future.

BELIEFS AND LEADERSHIP

The reason behaviour is dependent on prices is that a price movementin a particular direction represents the beliefs of one of the two crowdsand helps to fulfil the leadership function of that crowd. Because ourexperience of life through time is (apparently) linear and sequentialand (for most of us) our thought processes are of the same nature,there is a natural tendency to believe that what has just happened willalso continue to happen in the immediate future. If, for example,security prices have just risen, not only will the bulls feel satisfied, butthere will also be a tendency to assume that the rise will continue.There is therefore a clear ‘instruction’ to investors to buy more stock ifpossible. In a bullish environment, therefore, a rise in prices is the flagto which investors flock and pledge allegiance; in a bearish market afall in prices represents the harsh and terrible god to which cowedinvestors ultimately bend their knees.

INDIVIDUALS AS CROWD LEADERS

The role of price movements in the leadership function does notactually rule out the influence on crowd psychology of particular indi-viduals. Indeed, the emergence of a spokesman to publicize the beliefof the successful crowd greatly enhances the effect of recent pricemovements. From time to time, certain market traders or advisers dogain a reputation for ‘getting the market right’. Usually this reputationcomes either from making an accurate forecast (or ‘call’) of a marketwhen no one believed it or from sticking to a particular view of themarket trend longer than anyone else. At certain moments in a marketprice trend, sentiment may actually be influenced by such individuals.


However, there are two important points to be made about thepeople who provide this type of ‘leadership’. First, attention is givenonly to individuals who have already earned a reputation: very fewpeople will actually believe a new ‘guru’ the first time they make aforecast, no matter how accurate those forecasts subsequently turn outto be. Second, it becomes increasingly difficult for a particular indi-vidual to influence price movements on a regular basis. This apparentparadox occurs because, as was pointed out earlier, the more peoplewho believe in a trend, the fewer people there are left to perpetuate it.Sooner or later, a successful ‘call’ will be self-defeating because toomany people will believe in it. The price move generated by the callwill be quickly reversed and a large number of people who followedthe leader will actually lose money. Hard-won reputations, based on ahigh-profile exposure, are easily lost and can rarely be fully regained.

THE CONDITIONS FOR EFFECTIVELEADERSHIP

The identity of those individuals who are able to influence marketstherefore changes with time. Furthermore, this type of ‘leadership’ iscertainly not exercised as an ongoing influence: it is essentially a tran-sitory phenomenon that can only occur when market conditions areright. Specifically, there needs to be a state of tension between theopposing bullish and bearish crowds. Conditions of stress ensure thatthe reported comments or behaviour of a particular ‘leader’ can act asthe trigger for the large-scale desertion of investors from the unsuc-cessful crowd to the successful one. ‘Leadership’ exercised by indi-viduals can only be used within the context of the successful crowd.

INVESTMENT ADVISERS

Leadership of this nature needs to be differentiated from the regularday-to-day leadership provided by investment advisers to small groupsof investors. Essentially, investment advisers distil the plethora ofavailable economic, political and financial information and form aninvestment view that they dispense to their clients. They thereforeprovide positive leadership that, generally speaking, is acted on.

In relation to the phenomenon that we are analysing, however,investment advisers essentially lead subgroups within a bullish or


bearish crowd. Crowd theory certainly applies to each subgroup.Nevertheless, each subgroup will remain dominated by the main crowdto which it belongs. The strategy to which each subgroup respondsemotionally is the strategy of the ‘parent’ crowd. Successful adviserswill not, in fact, usually make recommendations that are completelyindependent of market conditions – indeed, they will themselves bepart of the crowd to a greater or lesser degree. However, even those rareindividuals who are capable of both recognizing market sentiment andof remaining aloof from it, and so are capable of providing contraryopinion advice, cannot fulfil the role of leader for the whole crowd.More precisely, it is because they remain aloof that they cannot bemarket crowd leaders. Only the happy minority will follow theiradvice, while the majority react either too slowly or not at all.

CONCLUSION

We have thus arrived at the following conclusions in this chapter. Thepursuit of trading profits as a common objective for investors ensuresthat stock market behaviour is a crowd phenomenon. The emotionalcommitment of individual investors to a particular set of trading posi-tions translates into an emotional commitment to a particular directionof price movement. Thus, each investor accepts a bull or bearargument, identifies strongly with others who have the same view andsubmits to the leadership provided by spokesmen (in so far as theyemerge) for their point of view.

The existence of two points of view ensures that conditions ofconflict will exist. Only one of the two points of view will, however,be successful and will reap the benefits of a price trend. Hence, thegradual emergence of a successful group in a financial market willensure that the conditions of conflict will intensify. The associatedstress works on the successful and unsuccessful groups in differentways. On the one hand, the successful crowd is encouraged to tightenits autonomy as well as intensify the pursuit of its strategic objectiveswith respect to prices. Consequently, an individual’s integrativetendency is stimulated to ensure that trading positions are marshalledfor maximum effect. Trading decisions therefore tend to becomeincreasingly non-rational as a price trend develops. As crowdmembers’ resources are thereby committed, the internal strength ofthe crowd transmutes itself into external strength. Propaganda againstthe losing crowd intensifies.


Obviously, the stress levels within the losing crowd will be intensi-fying. Eventually, members will respond by joining the successfulcrowd. New trading positions assist in fulfilling the successfulcrowd’s primary objective of ensuring that security prices movefavourably. Ultimately extremes of euphoria or pessimism occur, andthe conditions for a price reversal are created.

This analysis provides only the most basic framework for under-standing the behaviour of financial markets, but, even as far as it goes,it has a number of far-reaching implications. First, it provides a theo-retical framework that helps to explain both normal and abnormalbehaviour in financial markets. The same concepts can be used toexplain last week’s movements in share prices as well as the South SeaBubble of 1720 or the Wall Street Crash of 1929.

Second, it provides the theoretical basis for the dynamics ofbehaviour in financial markets that result in excesses of overtradingand create the conditions for price reversals. In other words, itexplains the internal mechanism by means of which the bull–bearcycle operates.

Third, it treats behaviour in financial markets as a naturalphenomenon that is subject to natural forces. (This raises some inter-esting possibilities as to the role of external influences – such as theweather – on crowd psychology.) The analysis has therefore moved along way from the simplistic view that price movements in financialmarkets are the wholly unpredictable outcome of wholly rational deci-sions made by completely independent people.

NOTES

1. This involves the purchase of a call option or a futures contract. There is animportant difference between the two. If held to expiry, an option is a ‘right’ tobuy or sell an underlying security that need not be exercised. However, if heldto expiry, a futures contract is a legal commitment to buy or sell an underlyingsecurity. Delivery of the underlying security can only be avoided if the futurescontract is closed off before delivery date.

2. This would involve either the purchase of a put option or the sale of a futurescontract. In some markets, it is also possible to sell stock ‘short’ – that is, to sellstock that is not actually owned.

3. Economic theory contends that savings essentially depend on the distributionof income over an expected lifetime. Hence, savings should be a largerpercentage of disposable income in the middle years of a person’s life than ineither the early or later years. See, for example, Farrell, M J (December 1959)


‘The new theories of the consumption function’, The Economic Journal, 69, pp 678–96.

4. For a description of the different concepts involved, see Tobin, James(February 1958) ‘Liquidity preference as behaviour towards risk’, Review ofEconomic Studies, 25, pp 65–86.

5. The usefulness of seeing the stock market in terms of three groups wassuggested to me by Alexander Elder, MD. See his book (1993) Trading for aLiving, John Wiley, New York. For simplicity, however, I prefer to include non-participants as part of the ‘unsuccessful’ group.

6. Changes are differences, and differences are items of information.7. This is a good example of what became known as ‘catastrophe theory’. See, for

example, Postle, Denis (1980) Catastrophe Theory, Fontana, London. Thistheory deals with ‘discontinuities’ in systems that otherwise change only grad-ually. The discontinuities allow a system to move to a new status without neces-sarily changing the structure of the system. Catastrophe theory postulates theexistence of only a limited number of discontinuity ‘types’and has therefore beenseen as providing a possible insight into evolutionary changes.


86

8

The shape of thebull–bear cycle

INTRODUCTION

We are now in a position to be able to look at the internal dynamics ofa complete bull–bear cycle in more detail. We have seen that theobjective of any crowd, taken as a whole, is to change the environmentfor its own ends. The objective of a crowd in a financial market is tochange prices in a particular direction: the bullish crowd will try toforce prices up, while the bearish crowd will try to force them down.Each crowd has its own life cycle. During periods when the market isin the process of reversing direction, the two crowds may coexist.However, once a trend has become established, only one crowd maydominate (see Figure 8.1).

Hence within a full cycle, first the bullish crowd will dominate, thenthere will be a period of uneasy coexistence with the bearish crowd,then finally the bearish crowd will dominate. This means that betweenany two given points in time, there will be a successful crowd and anunsuccessful one.

Individuals within one of these crowds will experience differentemotions from individuals in the other crowd. Members of thesuccessful crowd will be motivated by greed (or fear of missingfurther profits), will obtain pleasure from their success and will feelintegrated with like-minded investors. Members of the unsuccessfulcrowd, on the other hand, will experience fear (either of losing capital

The shape of the bull–bear cycle 87

or of not making any profits), will be prone to feelings of displeasureand will ultimately feel divorced from other members of the samecrowd. Eventually, the stress becomes unacceptable and members ofthe unsuccessful crowd desert to the successful crowd.

THE LIMIT CYCLE BETWEEN PRICES ANDSENTIMENT

The final arbiter of success, of course, is whether or not prices aremoving in the ‘right’ direction. We have already seen that price move-ments are the outcome of the conflict between the bulls and the bears,and that price trends (once started) tend to continue because pricemovements transmit information and contribute to the leadershipfunction. We have also seen that price reversals tend to occur when theinvestment community as a whole is more or less fully committed(financially) to one of the two points of view. Essentially, therefore,there is a limit cycle relationship between changes in prices andinvestor sentiment.

Integrativetendency

Bullish crowd Bearish crowd

Area ofcoexistence

Figure 8.1 The bull–bear life cycle

Investor sentiment has two aspects:

� expectations about the future trend in prices;

� the confidence with which these expectations are held.

Expectations obviously relate to whether the balance of sentiment isbullish or bearish, confidence to the willingness of market participantsto translate the expectations into actual transactions.1

Investor sentiment is reflected in such indicators as volume andmomentum. For the moment, we shall use a very simple proxy forsentiment – namely, a ratio of the number of bullish people to thenumber of bearish people. In Figure 8.2, this index is shown on thehorizontal axis. A move to the right reflects a relative increase in thenumbers expecting prices to rise, while a move to the left reflects arelative increase in the numbers expecting prices to fall. This index isdirectly influenced by both the direction and the size of the change inprices. The former has already been discussed. The latter is discussedbelow because crowds are more influenced by large price changes thanby small ones.

THE BIAS IN THE LIMIT CYCLE

Figure 8.2 shows the percentage change in prices against an index ofsentiment. Hence, the absolute price level will be rising above the zero


Figure 8.2 Stock market limit cycle

per cent line, and falling below the zero per cent line. As presented, thelimit cycle is biased to the right. This reflects an important fact, whichis particularly useful as an indicator of an imminent turning point –namely, that the confidence with which a view is being held usuallyturns prior to price reversals. Hence, just prior to market peaks,sentiment will begin to deteriorate as the percentage increase in pricesfalls. On the other hand, just prior to market troughs, sentiment beginsto improve as the percentage fall in prices decreases.

It is, in fact, possible to relate this rightward-biased price–sentimentlimit cycle directly to the relationships expressed in Figure 8.1. Figure8.3 therefore shows the bull and bear life cycles, the associatedperiods of rising and falling prices, and the limit cycle itself. All thatremains to be done is to convert the cycle of absolute price changesinto the cycle of percentage price changes. Turning points in thepercentage changes precede turning points in the absolute price level.

THE INFLUENCE OF ‘EXTERNAL’ FACTORS

The limit cycle between the prices of financial assets and sentimentdoes not, of course, exist in isolation: market prices are determined by


Inte

grat

ive

tend

ency

Abs

olut

e pr

ices

Cha

nge

inpr

ices

Figure 8.3 Limit cycle relationships

the physical supply of, and demand for, assets, which in turn aredependent on general financial conditions. In addition, sentiment infinancial markets is very dependent on the general mood that pervadesthe economic, social and political environment.

In particular, changes in financial markets reflect, and create, ageneral change in wealth within the whole community. Equity prices,for example, may rise because of improved economic circumstances,but economic circumstances may then be improved because of thewealth effect of higher equity prices. Quite simply, more people havemore capital that can be converted into current spending. Hence,rising stock prices become associated with optimism because theysuggest improved employment and income prospects for the future.However, falling stock prices create general feelings of pessimismbecause they have adverse implications for future employment andincomes. In this way, the emotions of one area of activity spill overinto another.

THE LIMIT CYCLE BETWEEN EQUITYMARKETS AND THE ECONOMY

We can express these interrelationships, if only in part, by means oflimit cycles that correlate financial asset prices with general economictrends. Such a correlation is shown in Figure 8.4. As it is usual to


Percentage risein equities

Percentage fallin equities

Troughin equities

Percentage fallin GNP

Peak inequities

Percentage risein GNP

Trough inGNP

Peak inGNP

Figure 8.4 Limit cycle relating equities to the economy

analyse economic activity in terms of changes, the limit cycle relatespercentage changes in gross national product (GNP) to percentagechanges in equity prices.2

The limit cycle could take one of a number of different shapes. Asshown in Figure 8.4, however, it is biased to the left. This bias reflectsthe important phenomenon to which we have already referred briefly,which is that security prices effectively anticipate the future byturning before fundamentals do. When the investment community as awhole is fully committed to a particular point of view, it can no longerrespond to news items that confirm that point of view. The emergenceof profit-taking after a price rise, or of ‘bear closing’ after a price fall,then creates the initial phases of a price reversal. Economic funda-mentals will turn at a later stage.

Limit cycles may actually be used to explore the relationshipsbetween any two economic variables you care to look at. It is apparentfrom the shape of the cycle in Figure 8.4 that there are four possibleinteractions between a pair of economic variables: two involving adirect relationship and two involving an inverse relationship.Economic theory normally only pays close attention to the former. Itfollows that the narrower the shape of the oval in the diagram is, thebetter will be the results of any correlation and regression analysis.This naturally ‘confirms’ the apparent validity of linear analysis, but,of course, ignores the dynamics of continuous negative feedback.

THE INFLUENCE OF SHOCKS

This analysis confirms that the price–sentiment limit cycle oper-ating in a financial market is also integrated with limit cyclesrelating that market to the wider economic, social and politicalenvironment. However, although accurate as far as it goes, theanalysis implicitly assumes that financial markets are in completeharmony with their environment and that, therefore, there isnothing new to learn about that environment. As we have alreadyseen, this assumption is ultimately incorrect:3 fluctuations that areintrinsic to a particular level of the hierarchy are inevitablymodified by fluctuations imposed from a higher level. Equity andbond prices therefore have to respond to new information about theenvironment at all levels of the hierarchy. In other words, they haveto respond to shocks.


Shocks occur because of a sudden divergence between current pricemovements and expected price movements and may derive from twosources:

� they may be triggered by an unexpected movement in prices them-selves;

� they may be precipitated by unexpected changes in the social,political or economic environment.

Furthermore, the shocks may be of either a pro-trend nature or acontra-trend nature.

PRO-TREND SHOCKS

The influence of pro-trend shocks can be very profound. For whateverreason, market participants suddenly find that expected prices arefurther away from the current levels than was originally thought. Pro-trend shocks are therefore usually sufficient to destroy the unsuc-cessful crowd and are always adequate to stimulate the integrativetendency of the successful crowd. Such shocks also involve a reducedlag between changes in sentiment and changes in prices, so that theassociated price movements are particularly dynamic.

Almost by definition, pro-trend shocks in a financial market materi-alize because of independent developments in the politico-economicenvironment. This implies that the shock has evolutionary implica-tions for the financial market and, therefore, has implications for thesubsequent pattern of price behaviour. We shall look at this in a littlemore detail in the next chapter. First, however, we need to look atcontra-trend shocks.

CONTRA-TREND SHOCKS ANDENERGY GAPS

At some (late) stage in a price trend, a financial market will neces-sarily experience a shock in the form of a contra-trend pricemovement. In principle, there are two ways in which this contra-trendmovement can arise:

� there may suddenly be an unexpected4 contra-trend movement inprices themselves;


� there may be an unexpected item of information concerningeconomic, social or political circumstances.

In either case, the successful crowd will experience a sudden diver-gence between actual and expected price movements. That crowd willthen start to revise its expectations concerning future price move-ments. Indeed, some investors will actually consider that the trend inprices has changed. This revision will inevitably weaken the inte-grative tendency of the successful crowd and thereby undermine theexisting price trend.

It is important to recognize, however, that price shocks occurringindependently of environmental data are more destructive than priceshocks occurring because of environmental data. This is true because,in the former case, the market will have entered an ‘energy gap’ thatcannot be bridged. In the latter case, however, pro-trend energy islikely to still exist and prices may quickly recover from the shock.

The role of energy gaps is, unfortunately, still not fully appreciated.It is directly related to the degree to which available financialresources are already committed to the prevailing trend. Once ‘all’ theliquidity is in the market, there is no energy left in the trend – a factthat explains the old market saying that prices peak on good news ortrough on bad. The last item of pro-trend news inevitably encouragesparticipants to commit themselves to the final series of trades thattakes the market into its terminal juncture. Genuine price reversalstherefore occur because of energy gaps.

Importantly, however, the price shock created by an energy gap isusually only the first part of a price reversal rather than the completereversal itself. The fact is that all natural systems respond to infor-mation shocks in a way that involves some retracement of the shock.There is an archetypal pattern of adjustment. As a result, there isusually a chance for knowledgeable investors to reverse their posi-tions in the market before it is too late. We shall now look at thispattern in some detail.

SHOCKS AND THE LIMIT CYCLE

As already described, the initial shock registers as a change in thedirection of price movement. Putting this another way, we can say thatthe shock occurs as the change in prices moves either from positive tonegative or from negative to positive.


Usually, however, it is possible to get a warning of an imminentprice shock from the behaviour of price momentum and investorsentiment. The relationship between either of these variables and thelevel of prices depends, of course, on the shape of the limit cycle.However, if we assume that the standard relationship between pricesand sentiment is as shown in Figure 8.2, then we can deduce thefollowing (see Figure 8.5).

If market prices are in a bull trend then, as that trend approaches itsterminal juncture, the rate of change in prices begins to drop. At thesame time, sentiment (however measured) will become less positive.Then the shock occurs as the change in prices moves through the zerochange line. This is shown by the movement from A to B in Figure 8.5.The relationship between prices and sentiment ‘jumps’ inwards, awayfrom the solution path defined by the limit cycle. Simultaneously, ofcourse, sentiment turns negative.

THE RETURN TO THE LIMIT CYCLE

As the limit cycle is essentially stable, however, behaviour will try toreturn to the solution path. If we continue to assume that prices are


Figure 8.5 The effect of a bearish information shock

entering a bear phase, then, essentially, the sequence of events is asshown in Figure 8.6. Initially prices may fall further after the shock,and this causes sentiment to do the same. However, the fall in pricesbegins to slow and, eventually, lower prices encourage ‘bear closing’.This, in turn, causes prices to rise. The rise in prices then stimulates areversal in sentiment.

The improvement in sentiment can be quite dramatic in relation tothe change in prices. Indeed, the strength behind the rise in prices mayeven be sufficient to take the market to a new high.5 In the absence ofother influences, however, the shock itself should be sufficient topreclude this.

Meanwhile, it is highly unlikely that either the rate of change inprices or the level of sentiment will achieve their previous highs.Eventually, higher prices encourage some profit-taking activity andprices begin to slip again. At first, sentiment may actually continue toimprove slightly as investors search for ‘cheap’ stock in anticipationof the ‘next’ rally. Unfortunately, however, the vast majority ofinvestors now hold stock and so their ability to buy any more is strictlylimited. Indeed, the subsequent price collapse is thereby assured. It isthis collapse that brings prices back to the solution path of the limitcycle.


Figure 8.6 The price–sentiment spiral after a peak

PRACTICAL IMPLICATIONS

There are three important conclusions that we can draw from thisanalysis:

� the adjustment process following a shock takes the form of a spiral;

� the price peak that occurs just prior to the shock is re-tested duringthe operation of the spiral – the re-test is ‘successful’ if marketprices move to a new high and ‘unsuccessful’ if they do not;

� sentiment appears to improve dramatically during the re-test, but,in most cases, falls short of that which was attained during theoriginal peak.

THE PATTERN OF ADJUSTMENT AFTERTROUGHS

Obviously, the whole analysis can be repeated for situations when amarket is turning from being bearish to bullish. This is shown inFigure 8.7. As the market comes to its reversal point, the rate of fall


Figure 8.7 The price–sentiment spiral after a trough

in prices will initially slow and sentiment will become less negative.Then the shock occurs as the change in prices turns positive. Thelocus of the price movement jumps upwards and inwards, awayfrom the limit cycle (from C to D in Figure 8.7). Then, the marketspirals back to the path of the limit cycle. During the spiral, absoluteprices are normally unlikely to make a new low, but neverthelessmay do so.

THE REVERSAL PROCESS

1. The shock

We have thus established an idealized model for the way that financialmarkets behave, and from this model we can isolate three distinctstages to a price reversal. In the first stage, the market reacts to beingeither overbought or oversold. As the market approaches its turningpoint, the vast majority of investors will, because of the integrativetendency, have the same (long-term) view about the future trend inprices. As the market begins to reverse, some individuals, who aretrading on relatively short time horizons, will be likely to developnegative (short-term) views and will start to neutralize all or part oftheir commitment to the market. Hence, an overbought market willencounter profit-taking on the part of investors and will fall, while anoversold market will encounter a bear squeeze and will rise. In bothcases, the market enters an energy gap.

There are, however, insufficient financial resources in the system tooffset the effects of short-term position-taking. What happens, therefore,is that prices adjust by a much larger amount than even the majority ofshort-term players are expecting. This adjustment therefore presentsitself as an information shock to the majority of market participants.

2. The re-test

The second stage of the reversal process essentially consists of the re-assertion of the sentiment of the previously successful crowd. Thefirst stage is seen as being technical rather than fundamental bymarket participants.6 It is, however, a trap for the unwary. Very often,the only way to recognize that this is part of a reversal pattern, ratherthan part of a fully fledged bull or bear move, is to analyse the actualbehaviour of the crowd in terms of underlying commitment to themarket (that is, to analyse their behaviour in terms of the volume of


transactions, their willingness to deal in the majority of securitiestrading in the market and their ability to open up new trading positionsand stimulate fast changes in prices). During this second stage of areversal pattern, these technical indicators of crowd behaviour willalmost certainly not confirm either the apparent dynamism of theprice movements or the sense of invincibility exhibited by traders.7

3. The new trend

The third and, in a sense, final stage of a reversal pattern is the actualchange in crowd sentiment as the fundamental news itself begins toalter. It is this stage that finally triggers an ever-increasing number ofpeople from the now unsuccessful crowd into changing their mindsand their trading positions.

The stress and tension of seeing fundamentals confirm the move inprices is finally reflected in an acceleration of buying (in a bullmarket) or of selling (in a bear market). This third stage, therefore,generates a sharp change both in price movements and in the volumeof transactions. It establishes the trend of the market, until a change infundamentals is once more on the horizon.

THE IDEALIZED THREE-STAGE REVERSALPATTERN

This analysis generates the concept of a three-stage reversal process.Idealized patterns are shown in Figure 8.8.

The important point to grasp is that the derived patterns are reflec-tions of the influence of a spiral. Stage 1 of the pattern (which may inpractice be more complicated than that shown in Figure 8.8) reflectsthe influence of a shock, and stages 2 and 3 reflect the spiral generatedby that shock.

Figure 8.8 highlights the relationship between actual price levelsand price changes. The lower part of the diagram makes explicit that:

� the rate of change in prices turns before the actual change;

� the reversal from an overbought or oversold condition creates ashock;

� re-tests of the initial reversal point are achieved at less extremerates of change in prices.


THE INFLUENCE OF FEAR

It is important to note the extent to which fear plays a role in thesepatterns. Fear is associated with feelings of isolation, inadequacy anduncertainty. It develops suddenly when expectations diverge fromactual events, and is only relieved by evasive action. In financialmarkets, therefore, fear persists until unsuccessful trading positionsare neutralized. Hence, as a market enters stage 1 of a reversal pattern,it is likely to be overextended, either because of fear of missing furtherprofits (in a bull market) or of making any losses (in a bear market).

There then follows the shock of an unexpected price change. Thiscreates an atmosphere of fear, but generates only a limited degree ofevasive action.

In stage 2, the re-test of the former peak or trough tends to assuagethe fears created by the shock, but always leaves a residue of intan-gible doubts.

However, in stage 3, prices move persistently against the beliefs ofthe majority until naked fear triggers a widespread neutralization(and/or reversal) of bad trading positions.


%ch

ange

inpr

ices

Shock

Shock

Market troughMarket peak

Figure 8.8 Idealized reversal patterns

It follows from this, of course, that the main indicators of investorsentiment (such as price momentum and volume) will show strongincreases during periods when fear is being heightened, and willsubside when fear is being reduced.

THE BIAS IN THE BULL–BEAR LIFE CYCLE

This analysis of fear implies that the disintegration phase of either abull or a bear life cycle will occur very quickly. The relevant life cycleis therefore biased to the right because the decline phase is mostclosely associated with fear. However, it is obvious that the fear of notmaking profits is of a different order of magnitude from the fear ofactually losing money. This implies, therefore, that the wholebull–bear cycle is biased to the right.

ASYMMETRIC INVESTMENT ATTITUDES

This rightward bias is assisted by another aspect of the effects of fear.The intensity of the fear of losing money means that most investorsactually have an asymmetric attitude to investment positions: theyprefer to hold stock rather than short positions.8 If something goeswrong with a purchase of stock, it is at least possible to regard it as along-term investment that will eventually come right. However, ifsomething goes wrong with the sale of something that is not evenowned, the short-term pressures to try again when the circumstancesare more favourable are irresistible. This whole process is, of course,made worse by the stock-based investment policies of the savingsinstitutions.

Bull markets therefore develop on the back of a gradual accumu-lation of stock, and a consequent reduction in liquidity. Furthermore,top patterns often take a long time to develop, because investors willcontinually take advantage of weaker prices to buy stock. However,when a bear market begins, not only do very few investors actuallyanticipate the fall, but there are insufficient bear positions to beclosed. There is therefore very little resistance to falling prices –indeed, liquidity can only be raised by selling, and this, by definition,forces prices down further. Bear phases therefore generally take ashorter period of time than do bull phases, and are very effective in


destroying wealth. Not surprisingly, bottom patterns tend to developquickly because so few investors actually have stock.

THE PRICE PULSE

We are now in a position to combine the two reversal patterns shownin Figure 8.8 above into one single pattern representing a completecycle. This is done in Figure 8.9. For simplicity, each wave of thepattern is given a different notation. Hence the initial upswing isdenoted wave 1, the subsequent re-test of the low is wave 2 and thebull impulse wave is wave 3. Then the initial drop from the high iswave A, the re-test of the high is wave B and the main bear run is waveC. This notation will be used in the remainder of the book.

The pattern is asymmetrical: stage 1 of the pattern is longer thanstage 2 because, as noted, bull markets generally take longer todevelop than do bear markets. From our knowledge of the operationof limit cycles, we know that the peaks and troughs will occur on aregular basis. This implies that the basic three-wave pattern of Figure8.9 will repeat itself continuously, thereby representing an archetypal‘heartbeat’ in a particular financial market. We propose to call this the‘price pulse’.


Figure 8.9 The price pulse

THE TIME HIERARCHY

It is necessary to register, however, that the price pulse is time frame-specific. That is, in any financial market, there will be a specific pricepulse that is applicable to each level of the market’s hierarchy of timeframes. The problem is determining what time frames constitutedifferent levels in the hiearchy.

This means that there will, in an ultimate sense, be a ‘mega-pulse’that will coincide with, and therefore trace out, the life cycle of thewhole financial market. The existence of such a pulse is the inevitablededuction from the theory of growth via evolutionary shocks.Eventually, therefore, there will be a financial market crash that willbe associated with a fundamental alteration in the currently acceptedmethods of production, distribution and exchange. At present,however, the concept is purely hypothetical and so is not meant to be aspecific forecast of the demise of the capitalist system. Its purpose ispurely to help illustrate the interdependence of the relationships thatwe have so far established.

PRICE–SENTIMENT LIMIT CYCLES

The concept of an identifiable life cycle for each and every financialmarket also implies the presence of identifiable metabolicprice–sentiment cycles. Once a life cycle has begun, it will trigger offa whole series of metabolic cycles in exactly the same way as the birthof a human being triggers off the rhythmic metabolic processes in thebody. Life cycles are separate concepts from metabolic cycles, butthey are dependent on one another. Hence, the movements of pricesover time will reflect the influence of both the life cycle and naturalprice–sentiment fluctuations.

LIMIT CYCLES AND THE TRANSMISSION OFSHOCKS

The important point is that the highest-level cycles have an impact atthe lower levels of the hierarchy. For periods, the lower-level cycleswill operate in harmony with the higher-level cycles. Gradually,however, divergences arise. Eventually, higher-level metabolic cycleswill transmit shocks to the lower levels because of the discrepancies.


These shocks create turning points. Each of them will create aspecific lower-level bull–bear life cycle and will then have its own inter-nalized metabolic rate. This process continues down through the hierar-chical levels until we reach the smallest of daily fluctuations. Hence,metabolic fluctuations at a higher level create life cycles at a lowerlevel. Simultaneously, however, the combination of all lower-levelcycles within the hierarchy constitute the ‘mega-pulse’ of the highest-level cycle. In this way, the system becomes completely integrated.

THE HIERARCHY OF FLUCTUATIONS

From these comments, we can make three deductions:

� the movements in prices will consist of both a hierarchy of meta-bolic cycles and a hierarchy of shock-induced fluctuations;

� we can confirm that all oscillations will harmonize with oneanother, both because the lower-level oscillations are triggeredby the highest-level oscillation and because the highest-leveloscillation is constructed from the effects of all the lower-leveloscillations;

� the patterns traced out by shock waves are reflected in the patternregistered by life cycles.

CONCLUSION

Let us now summarize the conclusions of this chapter by redrawingFigure 8.1 to include the greater detail that we have since uncovered(see Figure 8.10).

The top part of the diagram again consists of two life cycles – thatof the bullish crowd and that of the bearish crowd. Now, however, theearly stage of each life cycle is adjusted to reflect a three-stagereversal pattern, and each life cycle is biased to the right.

The second part of the diagram focuses attention on absolute prices.It incorporates the three-stage reversal pattern at both the bottom andthe top of market movements, and the bear phase is shorter than thebull phase.

The third part of the diagram converts the absolute price changesinto percentage price changes.



Inte

grat

ive

tend

ency

Abs

olut

epr

ices

Perc

enta

gech

ange

inpr

ices

Figure 8.10 The bull–bear cycle

The final part of the diagram shows the controlling influence of thelimit cycle. This cycle is the metabolic cycle that is ‘handed down’from a higher level of the hierarchy. As the rate of change in pricesturns from negative to positive, or from positive to negative, a shock isdelivered to the system. The result is the emergence of a lower-levelspiral adjustment at the point of inflection of the absolute price level.

We thus have a graphical representation of the basic features ofcrowd behaviour in financial markets. These are that:

� bull phases alternate with bear phases;

� each reversal pattern at the peak or trough of an oscillation consistsof three phases;

� rates of change in prices oscillate between upper and lower limits;

� rates of change in prices oscillate rhythmically over time.

We have already dealt in some detail with the fact that the bull andbear phases alternate with one another. We shall develop each of theother three features throughout the rest of this book. The next fewchapters will deal with the shape of the oscillations in more detail.Then we shall deal with the information provided by the technicalindicators of investor sentiment. Finally, we shall deal with the impli-cations of rhythmic oscillations.

NOTES

1. Confidence is, of course, an internal feeling. However, for market analysis, it isthe external signs of internal confidence that are important. For one thing,confidence may be reflected in a certain type of complacency where no moreactivity is being generated because participants are so certain they are right. Aswe shall see, this is a warning sign for markets.

2. Among other things, it is also possible to construct a limit cycle relating thebond market to economic activity. This implies the existence of a limit cyclebetween bonds and equities. Generally speaking, reversals in the bond markettend to lead reversals in the equity market by between three and six monthswithin the context of the 3¼-year business cycle.

3. It is valid only if it is applied to (relatively) high levels of the hierarchy over(relatively) short periods of time.

4. Unexpected, that is, by the majority.5. Mathematically, absolute prices will rise above the previous high if the rate of

change in prices above the zero per cent line is greater than the negative rate ofchange implicit in the shock.

6. In truth, it is essentially the other way round. The first stage of a reversalpattern is often the sign of changing fundamentals, while the second stage is atechnical response to fundamentals being discounted far too early.

7. This conclusion adds weight to the argument that market peaks do not neces-sarily occur at points of extremes in sentiment. Sometimes they do, but mostlythey do not. Usually, sentiment will deteriorate before prices start to fall. Theidea that market peaks occur at points of massive euphoria is probably createdfrom an illusion. During re-tests of market peaks, market traders are alwaysvery confident, particularly as the market moves into new territory. The markethas now finished ‘climbing its wall of worry’; it has had a correction, and hasthereby satisfied the fears of the traders; now the latter can relax in the face ofapparently improving fundamentals – even to the extent of buying intoweakness. The error, however, has been to mistake the change in sentiment


(which is quite noticeable) for the absolute level of sentiment (which remainsdepressed). Another way of saying this is that players still left in the marketregain their confidence without realizing that the total of the new resources thatare likely to be committed to the market has been reduced. The majority ofinvestors will therefore fail to recognize that the structure of the market ischanging, and it is usually this error that virtually guarantees the destructivenature of the subsequent bear market.

8. This is obviously not true of foreign exchange markets where a long position inone currency automatically involves a short position in another currency.


107

9

Energy gaps andpro-trend shocks

INTRODUCTION

It is quite clear from the analysis of Chapter 8 that each three-wavestage of the price pulse – whether up or down – is a response to thereceipt of information. As such, it reflects the process of learning. Thisis no accident. It has specifically been found that, when an individualis learning a new task, the ability to absorb and respond to the infor-mation passes through three distinct phases. This finding by HenryMills1 has not been given the credit it deserves, not least because itintroduces a significant qualification to the so-called ‘S’-shapedlearning curve. This latter was reputedly conceived by the Frenchsociologist Gabriel Tarde in 1890 and has been a favourite withacademics ever since. The two curves are shown in Figure 9.1.

Figure 9.1 The learning curve

The insight from Tarde’s curve is that it focuses on the three phasesof slow beginnings, accelerating growth and satiation. Its weakness,however, is that it ignores the causes of the slow beginnings, and failsto address the consequences of satiation. Mills’ findings specificallyaddress the first of these issues. The results of satiation are a separateconsideration.

In the initial phase of learning, some effort has to be applied toabsorbing and dealing with the new data. The mind has to concentrateand energy is focused. As a result, the ability to respond to the inflowof data actually begins to improve quite rapidly. Hence, if a new taskis being learnt, the trainee is increasingly able to do the task morequickly and with greater accuracy.

In the second phase, however, an interesting event occurs: the abilityto absorb new information suddenly wanes, the task of learning suddenlybecomes difficult and a sort of ‘forgetting’ occurs.2 The trainee knowsthat they can do the task, but somehow can’t quite do it effectively.

What happens is that the brain has to slow its learning process so thatenergy can be diverted to reorganizing the structure of memory. DonaldHebb3 proposed that the adjustment involves the transfer of informationfrom short-term memory to long-term memory. Short-term memorydepends on the electrical impulses of the brain, while long-termmemory depends on the structure of the neural network within the brain.In other words, persistent electrical activity eventually alters theconnections between brain cells so that something more than justtemporary learning takes place. It represents a paradigm shift.

The third phase, which follows this paradigm shift, is thereforecharacterized by an easy application of the learnt abilities. Thus, itdoes not reflect learning as such. Tasks can be carried out automati-cally with very little conscious concentration being applied. Short-term memory is released to cope with any other new information thatmight become available.

This move from an initial state of unconscious incompetence (theinformation shock), through the period of ‘forgetting’, to a final stateof unconscious competence (the use of learnt skills) is directlymirrored in the oscillation of the price pulse. Initially, there is an infor-mation shock, which is reflected in either a price squeeze on the bearsor a price fall against the bulls. Market prices move quickly as thecrowd starts to learn that the market is changing direction. Then therecomes a period of forgetting as the energy being directed towards thenew trend subsides. Finally, the new trend begins to assert itself inearnest, usually accompanied by appropriate economic data. By thisstage, of course, the crowd has effectively learnt that the trend isgenuine and is alert to the possibility of favourable economic data.


Therefore, there is a great willingness to commit money to the newtrend. This is the main impulse move – either up or down – as pricesmove along the path of least resistance.

ENERGY GAPS

The fact that the Mills learning curve shows the same three-waveprofile as is contained in the price pulse gives even more credence tothe latter. Furthermore, the logic behind both is essentially the same.What is not quite so clear, however, is the process whereby theupward-sloping learning curve converts into a downward-slopingcurve and then vice versa.

The answer lies in the concept of ‘energy gaps’. One type of energygap is actually implicit in the shape of the learning curve itself. Itoccurs during the period of ‘forgetting’ when internal energy isdiverted to restructuring memory. There is therefore a gap in theenergy being devoted to the emerging trend. This gap is caused by thelearning process and, hence, in a sense, is caused by the original‘shock’ of new information. Importantly, however, the energy gap hasan underlying support that links the initial information shock topermanent learning.

In a financial market, this ‘period of forgetting’ occurs when thecrowd denies the relevance of the information shock and accepts there-test of the high or low. This period corresponds to either wave 2 orwave B of the price pulse. Nevertheless, the crowd will internallyadjust its bias from bullish to bearish or from bearish to bullish even ifit doesn’t explicitly realize what is happening.4

There is, however, another type of energy gap. This gap actuallypresents a barrier to further progress and, quite simply, occurs as aresult of satiation. If an input of extra energy is not received at the gap,the system will fall into that gap. As a result, the system process willreverse itself. Hence, after a day’s work, both body and mind need anight’s rest. There is a natural energy gap, in the form of tiredness, thatconverts activity into rest. As a result, the active, creative processes ofthe day give way to the passive, recuperative processes of the night. Itis important, however, to recognize that sleep constitutes an activity inits own right: the activity is just of a different nature.5 At some stage,the activity of sleep will become satiated and create an energy gap.Hence, when a person has had sufficient sleep, they wake up. There is,thus, a simple, straightforward oscillation between activity and rest,mediated by energy gaps.

Energy gaps and pro-trend shocks 109

This analysis is simple, but profound. It points the way towards anunderstanding of all rhythmic fluctuations in terms of polarity swingsfrom one unbridged energy gap to another. Certainly, it can be appliedto the bull and bear swings in financial markets. Hence, it is apparentthat unbridged energy gaps create the contra-trend information shocksthat we analysed in Chapter 8. A market becomes overbought oroversold, such that no new resources can be committed to a trend andso a reversal begins. Unbridged energy gaps therefore create reversals(see Figure 9.2).

BRIDGING ENERGY GAPS

The only way that this rhythm of activity and rest can be altered is bybringing in energy from an external source to ‘bridge’ the gap. Forexample, excitement or stimulants may extend the activity of the day,while illness or sleeping pills may extend the passivity of the night.However, given the character of life as we understand it, these ‘exten-sions’ can only be temporary. Furthermore, the reaction to anextension can be quite dramatic: somebody who extends a workproject through the night is likely to find that the subsequent period ofsleep is longer than usual.


Figure 9.2 Energy gaps

Hence, an energy gap that is bridged will extend the trend. As wehave already observed, the first energy gap is, in fact, a reflection of abridge being formed. It is experienced during re-tests of highs andlows, when the market is in the process of learning that the direction ofenergy has changed. Such learning may be termed ‘first-order’learning, because the deep structure of the market remains unaltered.

On the other hand, however, the second energy gap is not automati-cally bridged. Such a bridge only occurs when a suitable externalshock informs the market that the quality of available energy haschanged. After this shock, the trend extends itself. The bridge is thus aform of learning and it can be classified as ‘second-order’ learningbecause the deep structure of the market alters. As already indicated inour activity/sleep example, extensions after the shock are likely to befollowed by a greater-than-usual reversal as the organism seeks toreturn to its normal pattern.

BRIDGING THE FIRST ENERGY GAP INFINANCIAL MARKETS

It is important to understand the underlying adjustments that take placeduring first-order learning, when the crowd is learning that a trendchange is occurring. The first point to make is that the adjustmentprocess is associated with a great deal of confusion. On the one hand,sentiment is still being influenced by the previous trend but, on the other,there may well be fundamental evidence that a change in trend isjustified. The integrative tendency of the new crowd is very weak at thisstage, while the integrative tendency of the old crowd is weakening.

The second point actually follows from this. It is quite apparent,during this first energy gap, that the market participants have a greatdeal of difficulty actually accepting a change in outlook.Consequently, commitments to both the old and new trends, takentogether, tend to drop. Uncertainty increases. Energy indicators – suchas momentum and turnover – tend to drop.

Underneath, however, the actual flow of investible funds is changingdirection as investors switch from one view of the market to anotherview. Thus it is that the first energy gap – the re-test of the high or low –provides the linking mechanism between the original contra-trendshock and the inevitable phase of ‘learnt response’ price movement.

What happens next is quite important. There will inevitably be apoint where the market itself registers that the trend has changed.


Obviously it is a very significant point for investors, although the wayit reveals itself is sometimes so subtle that it may be missed.Nevertheless, prices themselves signal that they are reversing awayfrom the first energy gap. Sometimes the reversal will occur on aspecific item of news and sometimes the reversal will just occur as aresult of a build-up of trading pressure. In either case, the signal islikely to be a greater-than-usual or unexpected price move, which maybe accompanied by an increase in the energy indicators such asvolume. The price move may be a break out of the downtrend thatformed while the energy gap was being negotiated, or it may be abreak out of the trading range that the energy gap has helped to form.In either case, however, the market will register the impact of an infor-mation shock (see Figure 9.3). We shall look at the implications ofinformation shocks in much more detail in Chapter 22.

BRIDGING THE SECOND ENERGY GAP INFINANCIAL MARKETS

As already mentioned, however, there is a limit to price movementsbased on ‘learnt responses’. In other words, there is a natural limit tothe amount of environmentally neutral growth that is possible.Inevitably, the limit cycle relationship between price movements andthe environment will begin to diverge. In principle, therefore, anenergy gap that is sufficiently large to form a barrier will arise. Theorganism or system, or (in this case) financial market crowd, will havebeen moving along the line of least resistance in relation to the envi-ronment. During the move, energy (or data) has been transferred from


Figure 9.3 Information shocks and the first energy gap

the environment to support the movement, but the market is notactually ‘learning’ anything new. Eventually, however, satiation setsin – the organism reaches the limits of this phase of its growth andfinds that old learnt responses no longer work.

In principle, this stage can be defined as the point of inflexion in thelife cycle of the system. After this point, the life cycle switches fromexpansion mode to contraction mode. In financial markets, this meansa switch from a bull phase to a bear phase (this was shown in Figure8.1 in the previous chapter).

Suppose, however, that an influx of external energy occurred that wassufficient to override the system’s predisposition to switch from growthto decline. This would mean a change in the quality of the life cycle. Infact, it would be represented graphically by a jump to a new life cycle(see Figure 9.4). This jump represents change in the form of revolution:the system metamorphoses to a higher level on the spiral of evolution.

However, it is important to recognize that the jump to a new life cyclerepresents a fundamental change in the quality of the system. As itstands, the diagram is drawn only in two dimensions, so it is all too easyto misinterpret the move from the lower life cycle to the higher one asjust involving a quantitative jump. In fact, what the diagram represents isan irreversible shift in the subjective meaning of the system: there is aninnovation, a paradigm shift, structural adjustment or shift in purpose.One of the clearest examples of the depth of change that is involved isthat of the fertilization of an ovum by a spermatozoon. Both the ovumand the spermatozoon have a specific life cycle: if they don’t meet, they


Shock

Figure 9.4 Evolutionary metamorphosis

both die; if they do meet, each jumps from its old life cycle to a new lifecycle. The latter, which is the formation of an embryo, includes but tran-scends both the ovum and the spermatozoon.

So, the complementary ideas of ‘transcending the old’ and of ‘irre-versibility’are the ones to keep in mind when using life cycle transfor-mations. Providing this is done, the concept is extremely powerful andcan be applied to all aspects of economic, financial and social devel-opment. In 1979, for example, the German economist GerhardMensch used the concept to demonstrate that an economy evolves bymeans of a series of intermittent impulses of innovation – a processthat he called ‘evolutionary metamorphosis’.6 In more recent years,the concept has been used by Charles Handy7 and Andrew Grove8 todemonstrate that, in order to avoid personal or corporate deterioration,there comes a stage when an evolutionary jump has to be made. To putthe same thing another way, if individuals or companies are notprepared to transcend old limitations, and are not prepared to adoptnew paradigms and new purposes, then a process of collapse sets in.9

THE IMPORTANCE OF TIMING

One point that is almost invariably missed in this type of analysis isthat there is only a very narrow time window in which the transfor-mation can occur. If that time window is missed, then either thechange cannot occur or the change is just too much and it becomesdestructive. In other words, evolutionary jumps can only occur atspecific points in an evolutionary process. This conclusion may bedifficult to accept, but the phenomenon can be clearly recognized in anumber of simple examples: children enter puberty at predeterminedtimes; water becomes steam at predetermined temperatures; and thequietude of winter becomes the vigorous stirring of spring during aspecific period in the relationship between the Earth and the Sun.

In fact, if we look a little more closely at the earlier example ofthe ovum and the spermatozoon, a further dimension of the myste-rious phenomenon of evolution reveals itself. This is that themoment of revolutionary change can only manifest under veryspecific conditions of synchronicity. That is, not only must theactive transforming energy (the spermatazoon) be able to makeexactly the right impact, but the passive receptive system (theovum) also has to be ‘ready’ to receive the energy. Under any othercircumstances, no change can occur.


How, then, can we determine when the recipient system is properlyreceptive to change? Nature, of course, has its own solution to thisparticular problem, but, in the context of our analysis here, the theo-retical answer to the question reveals one of the great secrets of lifeon this planet. A system becomes properly receptive only at the pointof inflexion of its life cycle. That is, it becomes properly receptive atan energy gap. Therefore, in Figure 9.4, the jump to a qualitativelydifferent life cycle effectively circumvents the collapse of the old lifecycle.

Note what will happen in other circumstances. If active energy isinput into a system prior to the point of inflexion, it might either berejected or create changes that are incompatible with the wider envi-ronment.10 On the other hand, if energy is input into a system after thepoint of inflexion, the system will already be reversing so that itscondition is inappropriate for the new energy.11

PRO-TREND SHOCKS IN FINANCIALMARKETS

We can now look in a little more detail at the impact of structural (orsecond-order) change in financial markets. The important elementsare shown in Figure 9.5, which traces out the initial phases of a bullmarket. The rally starts (11–21–31) with the first three waves of thelower-order price pulse (1–2–3). However, a pro-trend informationshock occurs that enables the market to jump the energy gap. As aresult, the market metamorphoses to a higher level (12–22–32), andexperiences the remaining quantitative movements of the lower-levelpulse (A–B–C) from a higher qualitative level. As presented, thelower-level price pulse continues to exert an influence on the marketpattern, even after the pro-trend shock. Consequently, the pro-trendshock merely causes the market to delay its downturn. However, thereare other alternatives. First, as already indicated, a pro-trend shock hasto be absorbed, and this involves some form of retrenchment, which,in theory, need not be perfectly aligned with a downturn in a lower-order price pulse. Second, if the pro-trend shock is particularlypowerful – which could well be the case if it emanates from a suffi-ciently high level – then the influence of the lower pulse may be all buteliminated. We shall revisit these issues in Chapter 22.

The corresponding pattern at the beginning of a bear move is shownin Figure 9.6. Here, the fall (A1–B1–C1) starts with the last three waves


of a lower-order price pulse (A–B–C). However, a pro-trend infor-mation shock forces the market over the energy gap that wouldotherwise turn prices higher. As a result, the market aligns itself with aqualitatively different structure, and experiences a further three waves(A2–B2–C2) down. The effectiveness of the shock is confirmed by thefact that wave 2 of the next cycle (which should be a re-test of atrough) makes a new low. This is a very common occurrence but, itneeds to be added, the timings need not automatically be determinedby the lower-order price pulse. On the one hand, some form of rallyafter the pro-trend shock is an inevitable consequence of the shockand may not necessarily tie neatly into wave 1 of the lower-level pricepulse. On the other hand, if the pro-trend information shock is signif-icant enough, the influence of the lower-level pulse may be washedaway. As already mentioned, we shall take another look at thesephenomena in Chapter 22.


Figure 9.5 Evolutionary metamorphosis during a bull move

ENERGY GAPS AND INFORMATION FLOWS

This analysis reveals the genuine importance of information at criticalmoments in the evolution of a price trend in a market. First, there isthe information shock of an unbridged energy gap. Then, during there-test of the reversal point, the market ‘learns’ that the trend is,indeed, changing. Some of the data that occur during this period willstrongly suggest that a reversal is justified. Finally, the market movesaway from the reversal pattern and follows the line of least resistanceup as far as the next energy gap. During this third phase, informationflows confirm the market trend.

During this whole process, it is apparent that the information flowsare either lagging behind market sentiment or are merely consistentwith it, but they are not leading market sentiment. Nevertheless, therewill come a point in the evolution of a trend when the market can go


Figure 9.6 Evolutionary metamorphosis during a bear move

no further on the basis of information that just confirms the trend.There is, for example, a limit to the number of people who can investsolely on the news that company earnings improved during the lastquarter. Unless the flow of news includes an item that is qualitativelydifferent, the market will reverse.

If such an item of information emerges, the second energy gap willbe bridged and the market will jump to a new structure. This item ofinformation will, in a sense, be ahead of the market. There may, forexample, be an official announcement that changes the responsibil-ities of the central bank. There may be an unexpected fiscalannouncement. There may be an ‘act of God’, such as an earthquakein a major financial centre, or there may be a drought or flood. Theremay be an outbreak of war. There may even be a small item of infor-mation that finally gives critical mass to a whole matrix of relatedinformation. The list is potentially infinite. However, there are criteriathat identify items of information that are pro-trend shocks:

� they are unexpected;

� they significantly alter the market crowd’s ‘mind map’ for thefuture – in other words, expectations are radically adjusted as aparadigm shift emerges;

� the shock is most likely to occur when the market is otherwisenaturally overextended and when price action may be flashingwarning signals for a reversal.

CHANGES IN THE QUALITY OFINFORMATION

This last point is extraordinarily important. It is usually (andcorrectly) argued that reacting to news can be dangerous. Specifically,it encourages traders to buy at high prices and sell at low prices,thereby creating an energy gap. Nevertheless, it is absolutely essentialto know when the quality of incoming data has changed. Awareness ofsuch changes can give an investor the critical edge. It enables them torespond with confidence to the news. Indeed, information shocks –and, hence, market extensions – actually occur when most marketparticipants least expect it. That is, the market is considered to have‘moved too far’. Under these circumstances, many short-term traderswill have taken profits and longer-term investors will not feelconfident about committing themselves further.12 This means thatfinancial resources will be available to respond to the shock.


IDENTIFYING THE SHOCK POINT

An information shock that bridges the second energy gap will almostcertainly impact during a very short period of time. As such, there willbe a ‘shock point’, definable in both price and time, that will be likelyto leave certain distinguishing features in the pattern of pricebehaviour and in indices of investor behaviour.

First, the shock point will be associated with a sharp movement inprices. This may occur either as a large expansion in the width of thetrading range for a given time period or as a gap in the price tradingrange between one period and another.

Second, momentum indicators (such as a rate of change indicator)will be likely to show signs of acceleration.

Third, investor volumes will jump and open interest in the futuresmarket will adjust sharply13 as the implications of the informationshock sink in.

Importantly, the information shock has implications for the gener-ation and measurement of cycles. It is part of the extraordinary math-ematical coherence of markets that the shock point can usually becalculated from the pattern of the price movement that follows theshock. We shall discuss this phenomenon in Chapter 22.

THE PATTERN OF A TREND

The bridging of the energy gap is probably the only time that datagenuinely ‘inform’ and lead the market. Furthermore, its occurrenceis important because the subsequent market price action is predictable– at some stage shortly after the information shock, a contra-trendmove will materialize as the cycle finally snaps in.

The most obvious conclusion from the patterns is that, following agenuine pro-trend shock, traders should buy weakness in a risingmarket or sell strength in a falling market. However, there are other,quite profound conclusions to be drawn from the analysis.

First, the presence of five waves in a trend – that is, three in thedirection of the trend and two against the trend – is a clear indicationof structural change. Five waves (or more) are the hallmark ofevolution (this is shown in Figure 9.7).

Second, the third wave within a five-wave movement is likely to bethe longest. This is because it consists of two parts – namely, theeventual movement along the path of least resistance after the contra-


trend information shock plus the initial response following the pro-trend shock.

Third, the fourth wave is unlikely to penetrate beyond the extremeof the first wave in the move. This is because the system has evolvedand there are constraints on the subsequent ability of the system toregress.

Fourth, the correction following the fifth wave is likely to begreater than the corrections that preceded it. This is because it emergesat a higher level of the market’s price–time hierarchy.

As many people now know, the five-wave pattern is the basis ofRalph Nelson Elliott’s ‘wave principle’. We shall therefore analyseElliott’s own work in a little more detail in Chapter 21.

CONCLUSION

Market oscillations are created by energy gaps. The first type ofenergy gap occurs where energy is being diverted to cope with theprocess of learning that a reversal has indeed taken place. The gap is,as it were, a ‘pause for thought’. Information becomes available that


Figure 9.7 The five-wave pattern of evolution

confirms the possibility that the reversal is genuine. However,sentiment is still coloured by the previous trend.

The partial retracement of the new trend – the re-test of the originalhigh or low – provides the platform for the subsequent impulse move.There is, therefore, a bridge across the energy gap that links the initiallearning situation to the impulse move. During this impulse move, themarket ‘knows’ that it is in a trend. All data will be understood in thiscontext. Data that do not confirm the trend will be dismissed orignored, while data that do confirm the trend contribute nothing newin terms of information. Prices will move along a line of leastresistance.

The second type of energy gap arises when a market is satiated oroverextended. In this case, no more resources can be committed to atrend. Consequently, the gap cannot be bridged and a reversal sets in.The reversal constitutes the contra-trend shock that was analysed inChapter 8.

The combination of energy gaps therefore creates a basic three-up,three-down oscillation in financial markets. This is the basis ofrhythmic cycles.

However, this basic wave pattern is subject to qualitative evolution.The second energy gap is both the point of satiation and the point atwhich a market is susceptible to revolutionary pro-trend shocks. Suchshocks have the ability to project the market into a new paradigm.They need to be recognized by market participants because theycreate a jump to a new life cycle. The emergence of such a life cyclecreates a market price pattern that can best be described as a five-wavepattern. Five waves are the hallmark of evolution.

NOTES

1. See Mills, Henry (1967) Teaching and Training, Macmillan, London.2. Mills, Henry (1967) Teaching and Training, Macmillan, London.3. Hebb, Donald (1949) The Organisation of Behaviour, John Wiley, New York.4. A similar phenomenon occurs following the intake of food. Some of the

existing energy has to be diverted to digest the food. As a result, the bodyinitially experiences a period of lethargy before recovering into a period ofhigher energy.

5. The fact that passivity is an active energy is usually missed and thereforecauses much analytical confusion. In financial markets, of course, the passivestage is a bear market. The larger bear markets destroy wealth rather than justgive everybody a period of rest and recuperation.


6. Mensch, Gerhard (1979) Stalemate in Technology, Ballinger, Cambridge,Massachusetts. Mensch was particularly forceful in arguing that depressionsoccur because of the absence of basic innovations. Although he didn’t specifi-cally say so, this is an energy gap.

7. Handy, Charles (1994) The Empty Raincoat, Hutchinson, London.8. Grove, Andrew S (1996) Only the Paranoid Survive, Doubleday, New York.9. We shall deal with the process of personal change in a little more detail in

Chapter 22.10. If, for example, an ovum is fertilized before it reaches the womb, severe

problems are created for the mother.11. An ovum cannot be fertilized when it is being expelled from the body during

menstruation.12. Obviously, the definition of short-term and long-term investment is highly

subjective.13. Open interest will eventually rise sharply. However, it may initially contract as

losing positions are unwound.


123

10

The spiral and thegolden ratio

INTRODUCTION

We have seen that the price adjustment process following an infor-mation shock can be represented by a spiral. The adjustment spiraloccurs over time, so it can be viewed in three dimensions. In afinancial market, these dimensions may be described as the rate ofchange in prices, the level of sentiment and time. As an example,Figure 10.1 reproduces the theoretical adjustment process during themovement from a bull market to a bear market. The left-hand part ofthe diagram represents the spiral relationship between the change inprices and sentiment, but with no ‘time’ dimension, while the right-hand part represents the corresponding ‘shock wave’ adjustment ofprices over time, but with no ‘sentiment’ dimension.

THE MATHEMATICS OF THE SPIRAL

From this it follows, of course, that price patterns should reveal thepresence of the spiral. We shall discuss this in more detail in Chapters11 and 12. However, there is another implication that is of profoundimportance. This is that price movements should be mathematicallyrelated to one another. The reason is that the spiral itself can bedefined using mathematics.


The source of all spirals is a geometric expansion of some form.Each number in a geometric expansion is obtained from the numberthat precedes it by multiplying it by a constant ratio. The most obviousis the so-called ‘doubling sequence’ where each term in the series isdouble that of its predecessor – namely,

2, 4, 8, 16, 32, 64, 128, and so on.

Here, the constant ratio is 2. It is readily apparent, even from thissimple example, that the higher the ratio, the more explosive will bethe result. It therefore follows that if a spiral is actually a ‘natural’phenomenon (in the literal sense of being a part of nature), the ratiomust be relatively small, although greater than unity.1

THE FIBONACCI NUMBER SEQUENCE

Of all the geometric expansions that can, and do, occur in nature, thereis one in particular that stands out above the others as being important.This is the Fibonacci sequence, which is based on the ratio 1.618. Thissequence is named after Leonardo of Pisa who, writing under thename of Fibonacci, published his famous Liber Abaci (or Book ofCalculations) in 1202. This book introduced the decimal system(which includes zero as the first digit in the sequence and which issometimes referred to as the Hindu–Arabic system) to Europe.Although Fibonacci was undoubtedly the greatest mathematician of

Perc

enta

gech

ange

inpr

ices

Perc

enta

gech

ange

inpr

ices

Figure 10.1 The price–sentiment spiral in two dimensions

The spiral and the golden ratio 125

the Middle Ages, it is perhaps ironic that he is remembered todaymainly because the nineteenth-century analyst Edouard Lucasattached his name to a sequence that appears in a trivial problem inLiber Abaci.

FIBONACCI’S RABBIT PROBLEM

The problem was presented in terms of the reproductive capabilitiesof rabbits – namely, how many pairs of rabbits could one pair producein a year?2 The first pair are allowed to reproduce in the first month,but subsequent pairs can only reproduce from their second monthonwards. Each birth consists of two rabbits. Assuming that none of therabbits dies, then a pair is born during the first month, so there are twopairs. During the second month, the first pair reproduces, creatinganother pair. During the third month, both the original pair and thefirst-born pair have produced new pairs. Consequently, there are threeadult pairs and two young pairs. If the analysis is continued, the resultsare as shown in Table 10.1, and it is apparent that the basic sequence(the Fibonacci sequence) is:

1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, and so on.

THE FIBONACCI SEQUENCE AND NATURE

Now, on the face of it, this sequence is of no interest to anyoneother than a student of mathematics or a rabbit breeder. However,

Table 10.1 Growth of a rabbit colony

Month Adult pairs Young pairs Total

1 1 1 22 2 1 33 3 2 54 5 3 85 8 5 136 13 8 217 21 13 348 34 21 559 55 34 89

10 89 55 14411 144 89 23312 233 144 377

mathematicians and scientists have discovered that the Fibonaccisequence can be found throughout nature, defining both theappearance of physical structures and the progress of change indynamic structures. Indeed, it appears that human beings find thosephenomena that are overtly related to the sequence intrinsicallypleasing, both to sight and hearing.3 Before pursuing these observa-tions any further, however, it is necessary to have a closer look at theproperties of the Fibonacci sequence.

THE PROPERTIES OF THE FIBONACCISEQUENCE

There are, in fact, three important properties of the sequence. The firstis that each term in the sequence (after the second) is the sum of thetwo terms that immediately preceded it. That is:

2 = 1 + 13 = 2 + 15 = 3 + 28 = 5 + 3

13 = 8 + 5and so on.

Such sequences – in which every term (after a certain point) can berepresented as a linear combination of preceding terms – are calledrecursive sequences. The Fibonacci sequence is the first-knownrecursive sequence.

The second important feature is that each term in the sequence,when divided by the term after it, approximates the ratio 0.618. To bemore precise, the ratio of successive terms oscillates around the limitof 0.618. The divergence from 0.618 is much greater for earlier valuesthan for later ones. The inverse of 0.618 is 1.618. Not surprisingly,therefore, the ratio of each term in the sequence, divided by the termbefore it, approximates 1.618.

The third feature of the sequence is that alternate terms are relatedto one another by the ratio 0.382 and by its inverse, 2.618. Hence, ifany term in the sequence is divided by the next-but-one number afterit, the result is 0.382, while if it is divided by the next-but-one beforeit, the result is 2.618. Again, the ratio is more accurate for calculationsapplied to later terms in the sequence than it is to earlier ones.

Finally, the same procedure can be repeated with numbers that lieincreasingly further away from one another. For example, numbers



that are three terms away from one another in the Fibonacci sequenceyield the ratio 0.236 and its inverse 4.236; numbers that are four termsaway from one another produce the ratio 0.146 and its inverse 6.853;and so forth.

THE IMPORTANT RATIOS

There are thus several ratios that can be derived from the Fibonaccisequence and various ways in which these ratios are related. Forexample:

0.618 × 0.618 = 0.3821.618 × 1.618 = 2.6182.618 × 1.618 = 4.236

1 – 0.618 = 0.3821.618 / 0.618 = 2.6180.618 / 1.618 = 0.382

0.382 × 0.382 = 0.146

It is apparent, however, that the two primary ratios are 0.618 and1.618. The others are essentially derivatives of these.

THE GOLDEN RATIO

The number 0.618 and its inverse, 1.618, are both known as the‘golden ratio’, and are usually denoted in the literature by the Greekletter phi (= �), which is the twenty-first letter of the Greek alphabet.It is of some significance that the golden ratio is functionally related to�5, which is equal to 2.236. Specifically:

0.618 = (�5 – 1) / 21.618 = (�5 + 1) / 2

THE GOLDEN RATIO IN GEOMETRY

The important role of 0.618 and its inverse, 1.618, becomes even moreexplicit when the idea is extended to geometry. Any straight line maybe divided in such a way that the ratio of the smaller part to the largerpart is equivalent to the ratio of the larger part to the whole. That ratiois always 0.618. This is shown in Figure 10.2, where

BC/AB = AB/AC = 0.618

In other words, the relationship between successive lengths – movingupwards from the smallest, through the largest, to the whole – isconstant. This means that, if the length AC is given as 1 unit, thensince AC = AB + BC,

AC = 0.382 + 0.618

And it also means that

0.382/0.618 = 0.618

The number 0.618 is the golden number, the associated ratio, 0.618:1,is the golden ratio, and the constant proportion between the lengths,0.382:0.618 = 0.618:1, is the golden proportion. The generic term forthe phenomenon is the golden measure.

THE GOLDEN MEASURE AND THEHUMAN BODY

Now, this is not just a rarefied exercise in mathematics or logic. Theimportance of the golden measure lies in the fact that it appears every-where in nature. For example, it can be found throughout the humanbody.4 The length from the base of a person’s foot to the navel is, onaverage, 61.8 per cent of that person’s total height. The distance fromthe navel to the middle of the neck is then 61.8 per cent of the distancefrom the navel to the top of the head; and the distance from the navel tothe knee is 61.8 per cent of the distance from the navel to the base of thefoot. Meanwhile, computer analysis of a so-called ‘ideal’ face, madefrom an amalgam of a large number of faces, confirms that the width ofthe head is 61.8 per cent of the height of the head; that the distance fromthe centre of the eyes to the tip of the nose is 61.8 per cent of thedistance from the eyes to the mouth; and that the distance from the tipof the nose to the chin is 61.8 per cent of the distance from the eyes tothe chin. Furthermore – and by no means finally – each bone of aperson’s finger is 61.8 per cent of the length of the adjoining bone.


Figure 10.2 The golden ratio

Of course, there will be significant individual divergences from allof these calculations; but the larger the number of individuals whosestructural proportions are measured, the more likely it will be that theaverage of the ratios will converge on the golden measure.

THE GOLDEN MEASURE IN NATURE

This is not the place to go into all the evidence for the presence of thegolden measure, but the point to be clear about is that it is ubiquitousin living systems. This raises very important questions, such as: Whyis this so? What does it imply about the laws of life? And should weactually be paying more attention to it? Unfortunately, althoughWestern culture recognizes the presence of the golden measure,5 itnevertheless tends to treat it as something of an accident, or evolu-tionary genetic outcome, and leave it at that. Very few take the time toponder some of the deeper implications.

However, this was not always the case. It is not known when thepresence of the golden measure was first recognized by humankind.There is certainly evidence that it was known about in the thirdmillennium BCE because it was incorporated into the building of thepyramids at Giza.6 However, what is clear is that, when one of the oldercivilizations came to know about it, it was treated with a particulardegree of reverence. Specifically it was treated, not as an outcome, butas an influence – not as an effect, but as a cause. It was considered to be,in some way, inherent in the blueprint of creation and therefore ‘prior to’the structures of this world. In short, the golden measure was consideredto be a direct reflection of the essential nature of God.

THREE TERMS FROM TWO

The mathematics of the golden measure give us some indication ofwhy this was considered to be the case. The main point is that thegolden ratio is the only ratio that allows the proportional relationshipsbetween three terms (AB, BC and AC in Figure 10.2) to be expressedby the relationship between two of those terms (AB and BC). In otherwords, if BC is to AB as AB is to AC, and if AC = AB + BC, then

AB/(AB + BC) = BC/AB = 0.618



THE GOLDEN RECTANGLE AND THEGOLDEN RATIO

Moreover – and this is important – the relationship stays constant, nomatter what absolute quantity is involved. That is, the goldenproportion operates in (and despite) the processes of growth anddecay. So, in Figure 10.2, AC can get bigger or smaller, but theinternal integrity of the relationships remains unchanged. A very basicdiagram can demonstrate this idea. In Figure 10.3, a rectangle ACDFis constructed on the line AC, derived from Figure 10.2. However, thedimensions of the rectangle are such that the width of the rectangle (ie,CD or BE) is equal to the length of the line AB. In other words, therectangle is drawn so that ABEF is a square.

THE SPIRAL OF RECTANGLES

Next, we can draw a square within the smaller rectangle BCDE, withone side on, and equal to, the line DE. See Figure 10.4. Two thingshave now happened. First, since by construction DE/EF = 0.618, thenthe perimeter length of the square DEHG is 61.8 per cent of that of the

Figure 10.3 The golden rectangle


square ABEF. Second, we have simultaneously created a rectangleGHBC. Since GH/CD = 0.618, then the perimeter length of therectangle GHBC is 61.8 per cent of that of the rectangle DEBC.

The process can then be repeated. See Figure 10.4 again. This timethe new square is drawn with one of its sides on, and equal to, GC. Therectangle GHBC then consists of a square GJIC and a rectangle JHBI.

THE GOLDEN SPIRALIt can immediately be seen that we are constructing a series ofcontracting squares, whose boundaries are touching in such a way thatthey appear to be spiralling towards some infinitesimally smallamount. Indeed, the points A, E, G, I, etc can be linked together by anappropriate spiralling arc. See Figure 10.5.

Logically, of course, we could have constructed the diagram differ-ently, drawing ever-increasing squares on the sides of rectangles. Theprocess would represent a spiral towards some infinitely large figure.

PROPERTIES OF THE GOLDEN SPIRAL

The golden spiral is a logarithmic spiral and, as such, it has twodistinctive features:

Figure 10.4 Golden rectangles


� it starts and ends in infinity – therefore it has no boundaries and itscentre can never actually be reached;

� it does not change shape – any straight line drawn from the centreintersects the spiral at the same angle (in Figure 10.6, therefore, theangle between the tangent at any given point and the radius drawnfrom the centre is a constant).

Because the golden spiral is defined by the ratio 0.618, and its inverse1.618, there are two further characteristics that are worth highlighting:

� each radius drawn from the theoretical centre of the spiral isrelated to the radius that precedes it at 90º by 1.618;

� each diameter of the spiral is related to the diameter that precedesit at 90º by the ratio 1.618.

THE GOLDEN MEASURE AND ANCIENTRELIGIOUS INSIGHT

We thus have a specific type of spiral – known as a logarithmic spiral– that is constructed using the ratio 0.618:1. It starts and ends ininfinity, and thereby identifies the unattainable point where the infi-nitely large is harmonized with the infinitesimally small. This,

Figure 10.5 The golden spiral

together with the other attributes of the golden measure that havealready been discussed, encouraged ancient religious authorities toconsider the validity of three inter-related propositions.

First, since the larger of the elements in equation 1 is equal to thesum of the other two elements, the golden number 0.618 was anexpression of the way that The Source of All Things divided itself intoThe Many. Or, to be even more esoteric, as the way that The Onedivided itself into Two. Specifically, the larger of the two portions wasthe passive, receptive, power; while the smaller portion was theactive, generative, power. Horizontally, this meant female and malerespectively; vertically, it meant spiritual and mental respectively. Butnote the extra element in this: the two lower-order (passive and active)polarities were balanced and harmonized by the higher-order whole.

Second, the proportional relationships made no claim about thecause of change. Change could be initiated from above (ie, fromthe whole), through either the active pole or the passive pole (ie, theparts); or it could be initiated from below by either the active part orthe passive part, and the whole adjusts. Since the mathematical rela-tionships required only that proportions remain constant, changeanywhere in the system was automatically supplemented by change inother parts. Among many other things, therefore, the golden ratio0.618:1 implied both that Man’s decisions were validated by God andthat God’s Will was implemented by Man.


Figure 10.6 Constant angles and the golden spiral

And third, because of the constancy of the relationships, the goldenproportion 0.382:0.618 = 0.618:1 was an expression of the fact thatchange can occur without involving chaos.

The golden measure was thus both the mathematical reconciliationbetween the Whole and Its Parts and the archetypal mathematicalformulation of the ancient saying, ‘As above, so below.’

THE ‘LAWS’ OF LIFE

On the face of it, these interpretations might be seen as extraordinarilyarcane and, accordingly, as having little or no relevance to the modernworld. But if we look a little deeper, there is a logic contained withinthem that not only bears consideration but also has a direct correspon-dence to modern scientific thought. Contemporary analysis is grad-ually coming to terms with three aspects of life on this planet. First,life is a self-organizing, constantly changing process that is stimulatedby (and therefore learns from) information ‘shocks’. Second, lifeinteracts with itself via circulation between opposite polarities. Andthird, life is attracted either to expansion or contraction. These arefundamental ‘laws’ for the existence of life; and there are no excep-tions. However, by definition, these laws have certain mathematicalrequirements, not the least of which is the need to reconcile randomfluctuations with stable processes. And the point is that the goldenmeasure may actually be part of the context within which this recon-ciliation can take place. Let’s just take a brief look at this.7

INFORMATION AND THE HUMAN MIND

The first area to look at is the human mind. The mind is a self-organizing system that learns by its interactions with the environment.Obviously, a part of this learning involves conscious reasoning. It isrelevant, therefore, that the concept of ratios is directly involved withthis process. Indeed, the words rational (‘of the reason’) and ratioci-nation (‘to reason’) are derived from the same Latin root as ratio. Oldercivilizations than ours had therefore obviously spotted the importanceof ratios to humankind’s ability to think clearly. To them the ability toreason meant the ability to use ratios. Suppose, therefore, that wewanted to present a ratio, or set of ratios, that simultaneously repre-sented the abilities of the human mind to perceive, reason and conclude.


What ratio(s) would we use? Put another way: what would be the char-acteristics of a ratio (or ratios) that best represented the various types ofinformation that a human mind can receive and generate?

RECOGNIZING INFORMATION

First, the human mind needs to be able to ‘recognize’ information.This means that it needs to be able to recognize a change, ordifference, between one state and another. This is the basis ofanalytical reasoning. In its simplest form, therefore, if we denote onestate as being A and the other as being B, and the human mind recog-nizes and responds to the difference between A and B, then the situ-ation could be represented by the ratio A:B. But note what happens ifA = B. There is no ‘difference’ as such, and the ratio A:B becomesequal to unity. Without difference, there is no information and noresponse. Hence, a ratio that represents analytical reasoning needs toconsist of unequal numbers so that it is not equal to unity.

UNDERSTANDING BY ANALOGY

Second, however, the ratio A:B does not in itself need human intelli-gence in order to create a reaction. The ratio could, for example,represent an amoeba’s response to a change in its environment. So asimple number is a necessary, but not sufficient, part of the represen-tation of human intelligence. In fact, human intelligence has a specificadditional dimension. This is the ability to understand by analogy – ie,by comparison. The ability to make comparisons facilitates attractionand repulsion. Consequently, a ratio that represents the human mindand its abilities would need to include three terms. That is, the appro-priate ratio would have to be able to compare not just A with B, butalso B with (say) C. Hence, the appropriate ratio would be somethinglike A:B = B:C. Providing that A and B are not equal, this allows forperception via both analysis and analogy. The ability to reason analyt-ically and analogically will ensure survival.

CREATIVE INSIGHT

Even so, there is a third element. The proportion A:B = B:C does notallow for the absolutely critical part of human intelligence that goes


beyond ensuring survival. This is the faculty of creative insight, orintuition, that encourages evolution. The very real problem, however,is that we do not really know where such insight (or intuition) comesfrom. It is one thing to collect pertinent facts, but it is yet another thingto claim that the critical insight based on those facts was purely theresult of analytical and analogical reasoning. Or, to put the same thinganother way, we can create the right quantitative conditions forinsights, but the essential ingredient of putting it all together and‘seeing’ a different qualitative dimension within a data set liessomehow beyond our ordinary everyday abilities.

Now, some of the greatest philosophers in the world – includingmore recent ones such as Alfred North Whitehead and William James– have suggested that intuition arises from a level of the mind’s organ-izing power that is not easily accessible. Specifically, they havesuggested that accessing intuition is very much a case of becomingpassive and allowing an insight to ‘drop through’ in one whole piece,as it were. The ‘effort’ therefore needs to go towards relaxing themind, rather than exciting it. If this is correct, then it suggests anintriguing possibility. This is that the mind is in a feedback loop withits own higher levels of organization.

SELF-ORGANIZING HIERARCHIES

This possibility is, in fact, not too different from the idea that wasinherent in religious attitudes to the golden measure. This is that thelower-order parts are organized by higher-order levels, but thathigher-order levels consist of lower-order parts. The differencebetween the two levels in this ‘hierarchy’ is that the whole is greaterthan the sum of its parts. And modern scientific thought nowabsolutely accepts the presence of such hierarchical organization inthe phenomenon of life. So, if we wanted to represent this hierar-chical, self-organizing aspect of the human mind, what set of ratioswould we use? Amazingly, the most appropriate set is the goldenproportion A:B = B:(A + B)! Indeed, it is only this set that allows forthe whole to consist explicitly only of its parts.

So, despite the apparently arcane origins of the golden measure, itprecisely contains within itself the important features of a self-organizing system. The golden measure recognizes information in theform of differences; it accommodates responsiveness to that infor-mation by allowing comparison; and it allows for feedback within a


system by recognizing the presence of hierarchical levels of self-organization. We are back to the three conditions for the existence oflife on this planet: self-organization stimulated by information, circu-lation between opposite polarities, and attraction to growth or decay.

These are profound conclusions that raise equally profound ques-tions about the insights of those who originally worked with thegolden measure. After all, only the terminology has changed. But, aswe have already seen in relation to the human body, the goldenmeasure does not appear to be just a metaphor. There seems to besomething quite practical about it.

METAPHOR AND REALITY

The first point to make is that metaphors and symbols are too oftentaken to be an irrelevance to ‘modern’ life. However, psychologistsare now very aware of the profound impact that metaphors andsymbols can have on the individual psyche – not only as representa-tions of inner realities, but also as triggers for inner change.Metaphors and symbols activate the right-hand side of the brain,which is non-linear and ‘holistic’. In fact, the right side of the brainhas to be harmonized with the left side for the faculty of creativeinsight to be enabled. It is therefore hubris to consider that the left-hand, linear side of the brain is more important, more creative andmore useful. And, as a corollary, it is immensely misguided toconsider that historically early insights into the nature of being wereincomplete or irrelevant. They may well have been complete: it is justthat they may not necessarily have been packed out with a moderndegree of scientific detail.8

The second point, however, is possibly more important. This is thatthe golden measure is not just a right-brain symbol. It is also clearlydefined by left-brain mathematical equations. The science of mathe-matics rightfully belongs to the realm of ideas, but its ultimate valuelies either in the accuracy of its description of known reality or in itsimplications for unknown reality. So, mathematics can, likemetaphors, point towards a truth that lies beyond the boundary ofknown experimental techniques. Hence, one inference that could bedrawn from the mathematics of the golden measure is that, if it isrelevant to life on earth, we really should be able to identify itspresence in indices that, in some way, reflect the behaviour of livingsystems. And it is arguable that one such system is a financial market.


FINANCIAL MARKET CROWDS

This idea holds water because the phenomenon of self-organization isuniversal. If every part of life has its own individuality but belongs toa larger group, then it follows that there are forces at work in humanbehaviour that originate from beyond each individual person. There isnothing esoteric about this (although that, too, is an option). All itmeans is that human beings come together in various ways and creategroupings whose influence is greater than the sum of their parts. At asimple level, there is ‘family’ or ‘tribe’; at a larger level, there is‘society’; and at a meta-level, there is ‘civilization’. And thesegroupings incorporate the influences of history, so self-organizationextends over time. The point is that these groupings are psychologicaland not necessarily physical. They are based on beliefs, under-standings and codes of behaviour that are transmitted from one personto another, and from one sub-group to another. And they are discrimi-natory: an individual is either in a particular overall grouping or out ofit. This is the basis of all forms of group competition and groupviolence. Individuals behave differently in groups than they do inisolation.

Within financial markets, the outcome of group behaviour is meas-urable by the fluctuation in prices. Indeed, financial markets thereforeform a wonderful laboratory for testing out ideas about groupbehaviour. Moreover, simple observation of market prices confirmsthat they oscillate – ie, they go up and down – so it is probable thatsome form of swing between opposite polarities is at work.Consequently, a practical working hypothesis includes three features:a) financial markets are a part of a wider socio-economic environmentand are therefore ‘organized’ by the latter; b) a market ‘organizes’ itsparts to conform with developments in the higher socio-economicenvironment; and c) the various hierarchical levels are linked by thetransfer of information that creates oscillations. Under thishypothesis, a group is a part of the natural order of things and, as such,its behaviour should reveal the influence of the golden measure.

THE GOLDEN MEASURE AND FINANCIALMARKETS

In fact, the presence of the golden measure in financial markets hasbeen known about since the early part of the last century. R N Elliott9


was the most celebrated exponent but others such as H M Gartley10

also utilized it very effectively. Elliott saw that corrections during amature trend would tend to remain above 61.8 per cent of that trend;both he and Gartley saw that a re-test of a high or low would belimited to 61.8 per cent of the initial breakaway from that high or low.To put the same thing another way, both saw that corrections weresomehow constrained by the golden proportion 0.382:0.618 =0.618:1.

One inference is that corrections are not random and chaotic eventsthat materialize out of nowhere. They somehow bear a natural rela-tionship to the cyclical process of which they are a part. We shallexplore this in more detail in the following chapters.

NOTES

1. If the ratio is less than unity, the series will consist of progressively smallernumbers.

2. Quoted in Vorob’ev, N N (1961) Fibonacci Numbers, Pergamon, New York.3. See, for example, the analysis by Huntley, H E (1970) The Divine Proportion,

Dover Publications, New York.4. The presence of the golden ratio in the human body is discussed in Ghyka,

Matila (1977) The Geometry of Art and Life, Dover Publications, New York.5. The recognition is also sub-conscious. It has been shown that if people are

presented with a large number of four-sided shapes ranging from a square to avery long, thin rectangle, most of them will choose the shape that correspondsto the golden rectangle. See Borissavlietch, M (1958) The Golden Number,Tiranti, London.

6. See, for example, Michell, John (1983) The New View Over Atlantis, Thamesand Hudson, London.

7. See, for example, Lawlor, Robert (1982) Sacred Geometry: Philosophy andpractice, Thames and Hudson, London.

8. Metaphors and symbols – particularly those that ‘work’ – point to great truths,even though they are not themselves the truth. Since the ultimate truth(whatever that may be) is – by definition – unchanged and unchanging, oldmetaphors and symbols may still be applicable, provided that we can under-stand them.

9. Elliott, Ralph N (1946) Nature’s Law: The secret of the universe, Elliott, NewYork, reprinted in Prechter, Robert (ed.) (1980) The Major Works of R N Elliott,New Classics Library, New York.

10. Gartley, H M (1935) Profits in the Stock Market, reprinted (1981) by Lambert-Gann, Pomeroy, Washington.


140

11

The mathematical basisof price movements

INTRODUCTION

It follows from the analysis of the previous chapter that the goldenratio is intricately linked to the growth of dynamic systems. As acrowd is a dynamic system, and financial markets exhibit crowdbehaviour, it follows that the golden ratio should be found in financialmarkets. So, the next question is: how does the ratio actually revealitself in financial markets?

We have already confirmed that all the points along a logarithmicspiral are mathematically related to one another. It therefore followsthat spiral adjustments in financial markets should exhibit the samerelationships. It will be remembered that when a financial market spiralis translated into two-dimensional space, the result is an ‘unstable’cycle. Each swing of this cycle is related to its predecessor by a loga-rithmic ratio. In the case of the golden spiral – to which we have paidspecial attention – the relevant ratio will be 1.618 and its derivatives.

In Figure 10.5 of the previous chapter, rectangles were used toconstruct a golden spiral.1 The important point is that the length ofeach rectangle measures a ‘width’ of the spiral, and that alternaterectangles reflect opposite ‘swings’ of the spiral. Hence, if we transferthe spiral into the two dimensions of price and time, it follows thatsuccessive swings will be related to one another by the ratio 2.618:1.2This relationship is shown in Figure 11.1, where:

CD/BC = 2.618

The mathematical basis of price movements 141

THE CALCULATION OF PRICE TARGETS

This implies that if we can identify the presence of the ‘unstable’cyclein price movements, we should also be able to calculate precise pricetargets from the number 2.618. Initially, the market responds posi-tively to a shock (which may be suggestive of either a new trend or aresumption of an old one). It then reverses direction under theinfluence of the spiral mechanism. Finally, it jumps in a dynamicmove – the extent of which is determined by the ratio 2.618:1. In otherwords, the target level for the thrust is 2.618 times the arithmeticallength of the last wave of the base or top pattern that preceded it.Hence, in Figure 11.2 the target price (Pt) is given as:

Pt = P2 + (P1 – P2) × 2.618for rising markets, and:

Pt = P2 – (P2 – P1) × 2.618for falling markets.

Despite the general accuracy of this calculation, it is worth noting thatthe mathematics sometimes need to be conducted in percentage terms.This is particularly true of bond markets, which are particularlysusceptible to the psychological influence of percentage changes.Here, the relevant formulae are:

Figure 11.1 The golden ratio in price fluctuations


Pt = P2 × (1 + (1 – (P2/P1))) × 2.618

for rising markets, and: Pt = P2 × (1 + (P1/P2) – 1)) × 2.618

for falling markets.

It is a simple exercise to conduct a few calculations on historical datato see if a particular market is responding to percentage changes or toabsolute changes.

THE APPLICATION OF THE TARGETFORMULA

One of the most important demonstrations of the target formula datesback to the Wall Street Crash in the United States. Between the stockmarket high in 1929 and the low in 1932, the Dow Jones IndustrialAverage Index collapsed from 386 to 42 – a fall of 344 points.Multiplying 344 by 2.618 gave a targeted rise of 901 points. This meantan Index target of 942, just 58 points short of the psychologicallyimportant 1000 level (see Figure 11.3). This target was achieved almost34 years3 later in 1965–66, after which followed a multi-year correction.In fact, the level of 1000 was not seriously penetrated until early 1983.

However, the 2.618 ‘spiral’ formula works in falling equity marketsas well. Figure 11.4 shows the chart for the Hong Kong stock market,as represented by the Hang Seng Index. During the spring and earlysummer in 1997, the Hang Seng rallied to an intra-day high of

Rising market Falling market

Figure 11.2 The 2.618 calculation

16820.31 on 7 August. There then followed a sharp 23.3 per cent dropinto early September. The next move, however, was a standard re-testof the high. The rally amounted to 2342.84 index points (that is, 18.16per cent). Accordingly, when the next phase of the sell-off began in thewake of the Asian liquidity crisis, the projected fall was one of6133.56 points, to an index level of 9109. This was a fall of 47.5 percent (or a total fall of 54 per cent from the August high). On 27 October 1997, the market closed at 9059.89, having beensomewhat lower during the day.

0

200

400

600

800

1000

1200

1400

Oct

-28

Oct

-30

Oct

-32

Oct

-34

Oct

-36

Oct

-38

Oct

-40

Oct

-42

Oct

-44

Oct

-46

Oct

-48

Oct

-50

Oct

-52

Oct

-54

Oct

-56

Oct

-58

Oct

-60

Oct

-62

Oct

-64

Oct

-66

Oct

-68

Oct

-70

Oct

-72

Oct

-74

Oct

-76

Oct

-78

Oct

-80

Oct

-82

Inde

x le

vel

Dow Jones Industrial Average(Weekly, Friday close)Data source: Economagic

261.8% of 1929-32 Crash = 942

Figure 11.3 The Dow, 1928 to 1982

Figure 11.4 The Hang Seng stock market crash, 1997


7,000

9,000

11,000

13,000

15,000

17,000

19,000

Mar

-97

Mar

-97

Apr

-97

Apr

-97

May

-97

May

-97

May

-97

Jun-

97

Jun-

97

Jul-9

7

Jul-9

7

Aug

-97

Aug

-97

Sep

-97

Sep

-97

Oct

-97

Oct

-97

Oct

-97

Nov

-97

Nov

-97

Dec

-97

Dec

-97

Jan-

98

Jan-

98

Feb-

98

Inde

x le

vel

Hang Seng Index(Daily, high, low, close)Data source: Economagic

261.8% of Sep-Oct rally = 9,109


EXAMPLES FROM THE UK GILT-EDGEDMARKET

We shall now look in some detail at government bond markets. Theyare highly liquid markets in the sense that willing buyers and sellerscan be matched in large numbers within a relatively narrow range ofprices, and they are efficient in so far as information is transmittedvery quickly. Furthermore, bond markets are particularly applicable tothe subject matter of this book because they are also very responsive toa wide range of economic, political and social influences.

Bond markets are also useful markets to monitor for purely practicalreasons. They can be very simply represented by a general index, bythe price of a single bond or by the yield of a single bond. First, apartfrom slight nuances with regard to coupons, one treasury bond is muchthe same as any other within a particular maturity range. Second, in allfixed interest markets, yields are inversely related to prices. Therefore,a chart history of any bond price over any given period would not onlybe representative of a whole range of stocks within a market, butwould also be a mirror image of a chart history of yields over the sameperiod.

Hence, Figure 11.5 shows the price movements of the UK treasurybond market (called the ‘gilt-edged’ market) between 1970 and 1994.Here, the gilt market is represented by an ultra-long-dated stock,Treasury 2½ per cent, redeemable on or after 1 April 1975.4 Thisparticular stock has been chosen because it provides a long and

10

15

20

25

30

35

40

Jan-

69

Jan-

70

Jan-

71

Jan-

72

Jan-

73

Jan-

74

Jan-

75

Jan-

76

Jan-

77

Jan-

78

Jan-

79

Jan-

80

Jan-

81

Jan-

82

Jan-

83

Jan-

84

Jan-

85

Jan-

86

Jan-

87

Jan-

88

Jan-

89

Jan-

90

Jan-

91

Jan-

92

Jan-

93

Jan-

94

Pric

e le

vel

UK Treasury 2 1/2% ‘Irredeemable’(Monthly closing prices) Data source: Datastream

Figure 11.5 The UK gilt-edged market, 1970 to 1994

consistent history, without suffering distortions provided by the ‘pullto redemption’.5 The chart shows the massive collapse in pricesbetween January 1972 and January 1975, followed by the volatile risefrom 1975 to the high in January 1994. It also shows the so-called‘crash’of 1994. Roughly, we can identify three major bull phases overthe period: October 1976 to January 1978, September 1981 toNovember 1982, and April 1990 to January 1994. Importantly, andpersuasively, each of these bull phases – measured in percentageterms – was 2.618 times the correction that immediately preceded it.6

The years of 1972, 1973 and 1974 were dark years for both the giltmarket and the UK equity market. Inflation was rising rapidly and thegovernment boosted the economy when it should have been pulling inthe reins. Between January 1975 and January 1976, however, therewas a sharp rally in asset markets. In the case of Treasury 2½ per cent,the rally was one of 33 per cent. This rally was sufficient to convinceeven the most committed bears that a change in trend had taken place.However, between January 1976 and late 1976, when the IMF becameinvolved with the state of Britain’s finances following a sterling crisis,the market reversed. Treasury 2½ per cent fell by 20.13 per cent. If thespiral hypothesis is correct, then the subsequent rally should haveamounted to 52.7 per cent (that is, 20.13 × 2.618), giving a target of23.47. Between the end of October 1976 and the major peak in January1978, the market rallied by 54.5 per cent to bring the price of Treasury2½ per cent to 23.75 (see Figure 11.6).

There then followed a lengthy, and frequently interrupted, bearmarket down into September 1981. The last phase of this fall – from


Figure 11.6 The UK gilt-edged market, January 1974 to October 1979

14

16

18

20

22

24

26

Jan-

74

Apr

-74

Jul-7

4

Oct

-74

Jan-

75

Apr

-75

Jul-7

5

Oct

-75

Jan-

76

Apr

-76

Jul-7

6

Oct

-76

Jan-

77

Apr

-77

Jul-7

7

Oct

-77

Jan-

78

Apr

-78

Jul-7

8

Oct

-78

Jan-

79

Apr

-79

Jul-7

9

Oct

-79

Pric

e le

vel

UK Treasury 2 1/2% ‘Irredeemable’(Monthly closing prices) Data source: Datastream

261.8% ofJan - Oct 76 fall

= 23.75


July 1980 to September 1981 – was a steep decline of 23.44 per cent(see Figure 11.7). The calculated target after this low involved a rallyof 61.4 per cent to a price of 27.03. The actual close on 2 November1982 was 27.00.

The following six and a half years were very volatile, to say the least.In particular, they covered the period leading up to and following theequity crash of 1987. The price of Treasury 2½ per cent rallied into ahigh (of 30.0) in April 1986, fell (to 24.63) prior to the equity crash, thenrose into a re-test high (of 29.50) in March 1989. However, betweenMarch 1989 and April 1990, the bond fell by 29.5 per cent (to 20.81),taking the total fall since the high of April 1986 to 30.63 per cent (seeFigure 11.8). The calculated upside target was 37.5, representing a riseof 80.2 per cent. The actual high in January 1994 was 37.63. The subse-quent fall is now referred to as the bond ‘crash of 1994’.

ADDITIONAL EXAMPLES

Before moving on to the US Treasury market, it is perhaps just worthcommenting a little further on two specific episodes shown in Figure11.5. These are the fall in gilt prices between January 1972 andJanuary 1975 and the sharp rally into the high of April 1986. For thisanalysis, we shall use a different gilt stock – namely, Treasury 5½ percent 2008/12. This stock has been chosen in order to highlight the factthat price targets may be calculated in arithmetical terms as well as inpercentage terms.

16

18

20

22

24

26

28

Jan-

79

Mar

-79

May

-79

Jul-7

9

Sep

-79

Nov

-79

Jan-

80

Mar

-80

May

-80

Jul-8

0

Sep

-80

Nov

-80

Jan-

81

Mar

-81

May

-81

Jul-8

1

Sep

-81

Nov

-81

Jan-

82

Mar

-82

May

-82

Jul-8

2

Sep

-82

Nov

-82

Jan-

83

Mar

-83

May

-83

Jul-8

3

Sep

-83

Nov

-83

Inde

x le

vel

UK Treasury 2 1/2% ‘Irredeemable’(Weekly, Friday close) Data source: Datastream

261.8% ofJul 80 to Sep 81 fall

= 27.03

Figure 11.7 The UK gilt-edged market, January 1979 to November 1983

In the early 1970s, Chancellor Barber (under Prime Minister Heath)injected a massive dose of inflation into the UK economy. Long bondswere very badly hurt. Between January 1972 and January 1975, the priceof Treasury 5½ per cent 2008/12 fell by 40.37 points – that is, from aprice of 73.25 to a price of 32.88 (see Figure 11.9). This was a fall of 55.1per cent. In the price action prior to the collapse, Treasury 5½ per cent2008/12 rallied by 15.25 points. Using the arithmetical formula, thetargeted fall was 33.33 (that is, 73.25 – (2.618 × 15.25)). The differencebetween forecast and outcome was therefore just over 1 per cent. Theincredible implication of this calculation is that, although the fall wasdirectly attributable to the policy errors made by Heath and Barber, itwas actually foreshadowed by market behaviour in the period 1969–71.

Between February 1985 and April 1986, the gilt-edged market had asharp rally into a peak that was not subsequently exceeded until 1993.The rally consisted of an initial base, lasting from February 1985 toFebruary 1986, followed by a sharp impulse wave. Within this, the lastphase of the base pattern between 29 July 1985 and 31 January 1986was a fall of 5.25 points. Using the arithmetical formula, the projectedrally for Treasury 5½ per cent 2008/12 was for a rise of 13.75 points toa level of 71.75. This level was hit precisely on 17 April 1986 (seeFigure 11.10).

These calculations are very compelling. Critics may justifiably arguethat not all of them are absolutely precise to the nearest two decimalplaces. However, the fact is that the forecasting errors are normally verysmall in relation to the huge moves that are being analysed.


19

21

23

25

27

29

31

33

35

37

39

41

Jan-

86

May

-86

Sep

-86

Jan-

87

May

-87

Sep

-87

Jan-

88

May

-88

Sep

-88

Jan-

89

May

-89

Sep

-89

Jan-

90

May

-90

Sep

-90

Jan-

91

May

-91

Sep

-91

Jan-

92

May

-92

Sep

-92

Jan-

93

May

-93

Sep

-93

Jan-

94

May

-94

Sep

-94

Inde

x le

vel

UK Treasury 2 1/2% ‘Irredeemable’(Weekly, Friday close) Data source: Datastream

261.8% of Apr 86 to Apr 90 fall = 37.5

Figure 11.8 The UK gilt-edged market, January 1986 to September 1994


J A S O N D J F M A M J J A S O N D J F M A

1984 1985 198654

56

58

60

62

64

66

68

70

72

Treasury 5 1/2% 08/12Source: Datastream

Figure 11.10 The UK gilt-edged market, July 1984 to April 1986

1970 1971 1972 1973 1974 197530

35

40

45

50

55

60

65

70

75

Treasury 5 1/2% 08/12Source: Datastream

Figure 11.9 The UK gilt-edged market, January 1970 to January 1975

EXAMPLES FROM THE US TREASURY BONDMARKET

These findings can be confirmed by an analysis of the US Treasurybond market. Figure 11.11 shows an index of Treasury bond prices,calculated by deducting the constant maturity yield on 20-year bondsfrom 100. The period covered in Figure 11.11 runs from the secularlow of 1981 to just prior to the Equity Crash of October 1987. Withinthis pattern, we shall be looking at price behaviour in the autumn of1981, then from late 1981 to late 1982, and finally from the summer of1983 to the spring of 1987.

Figure 11.12 shows the dying phase of the great inflation-generatedbear market of the 1970s and early 1980s. In October 1981, the marketrallied in early October, and then fell to make a final – albeit marginal– new low. The last fall on the 30-year index was 1.1 per cent.Multiplying this by 2.618 indicated a rally of 2.9 per cent, or an indextarget of 87.25. The index closed at 87.24 on 27 November.

The market then fell steeply, as it re-tested the October 1981 low. ByFebruary 1982, the index had fallen by 2.4 per cent. However, the basepattern continued for some months and was not finally terminated untilJune 1982 (see Figure 11.13). When the market broke out of the basepattern in August 1982, it was possible to anticipate a total rally from theJune turning point of 6.27 per cent. This corresponded to an index targetof 89.62. The market peaked at an index level of 89.65 on 19 November1982, formed a top pattern, and collapsed again into May 1984.

84

86

88

90

92

94

96

Jan-

81

Jul-8

1

Jan-

82

Jul-8

2

Jan-

83

Jul-8

3

Jan-

84

Jul-8

4

Jan-

85

Jul-8

5

Jan-

86

Jul-8

6

Jan-

87

Jul-8

7

Jan-

88

Jul-8

8

Jan-

89

Jul-8

9

Jan-

90

Jul-9

0

Jan-

91

Jul-9

1

Jan-

92

Jul-9

2

Jan-

93

Jul-9

3

Jan-

94

Jul-9

4

Jan-

95

Jul-9

5

Jan-

96

Jul-9

6

Jan-

97

Jul-9

7

Pric

e in

dex

US 30-Year Treasury Bonds(100 minus constant maturity yield, weekly)

Data source: Federal Reserve Board

Figure 11.11 US Treasury bonds 1981 to 1987



Figure 11.14 draws attention to the last phase of this fall. BetweenOctober 1983 and May 1984, the index fell by 2.56 index points or by2.89 per cent. Applying the spiral formula to these two figures, theformer predicted a rise of 6.70 index points to 92.76, while the latteranticipated a rise of 7.56 per cent to 92.57. The market peaked at92.84 in April 1986.

84.5

85

85.5

86

86.5

87

87.5

88

01-S

ep-8

1

08-S

ep-8

1

15-S

ep-8

1

22-S

ep-8

1

29-S

ep-8

1

06-O

ct-8

1

13-O

ct-8

1

20-O

ct-8

1

27-O

ct-8

1

03-N

ov-8

1

10-N

ov-8

1

17-N

ov-8

1

24-N

ov-8

1

01-D

ec-8

1

08-D

ec-8

1

15-D

ec-8

1

22-D

ec-8

1

29-D

ec-8

1

Inde

x le

vel

US 30-Year Treasury Bonds(100-constant maturity yield, daily)Data source: Federal Reserve Board

261.8% ofOctober 1981 fall

= 87.25

Figure 11.12 US Treasury bonds, October to November 1981

Figure 11.13 US Treasury bonds, November 1981 to November 1982

84

85

86

87

88

89

90

91

Jul-8

1

Aug

-81

Sep

-81

Oct

-81

Nov

-81

Dec

-81

Jan-

82

Feb-

82

Mar

-82

Apr

-82

May

-82

Jun-

82

Jul-8

2

Aug

-82

Sep

-82

Oct

-82

Nov

-82

Dec

-82

Jan-

83

Feb-

83

Mar

-83

Apr

-83

May

-83

Jun-

83

Jul-8

3

Inde

x le

vel


261.8% ofNov 1981 to Jun 82 fall

= 89.62

In this last example, there is obviously a slight difference in accuracybetween the two targets: the arithmetic calculation was more precisethan was the percentage calculation. In fact, it needs to be remem-bered that the implied fall in long-dated bond yields between May1984 and April 1986 translates into a massive rise in actual bondprices, and that the resulting error is actually quite small for thosewho played the long side. Nevertheless, the difference obviouslyhighlights the advantages of conducting the calculations in both arith-metical and percentage terms. It also confirms that some leeway hasto be allowed in generating targets, particularly if the market is accel-erating sharply (as was the case with US bonds in 1986). Targets arelikely to be temporarily penetrated as traders panic but this, in itself,creates the conditions for a reversal and underscores the effectivenessof the target calculations.

CONCLUSION

We have looked closely at price swings in two of the world’s majorbond markets over relatively long periods of time. We have alsopicked out two important examples from equity markets. There isstrong evidence that the main impulse wave of either a bull market orbear market is related to the base or top pattern that preceded it by the


Figure 11.14 US Treasury bonds, October 1983 to April 1986

84

85

86

87

88

89

90

91

92

93

94

Jul-8

1

Oct

-81

Jan-

82

Apr

-82

Jul-8

2

Oct

-82

Jan-

83

Apr

-83

Jul-8

3

Oct

-83

Jan-

84

Apr

-84

Jul-8

4

Oct

-84

Jan-

85

Apr

-85

Jul-8

5

Oct

-85

Jan-

86

Apr

-86

Jul-8

6

Oct

-86

Jan-

87

Apr

-87

Jul-8

7

Oct

-87

Inde

x le

vel


261.8% ofNov 1981 to Jun 82 fall

= 92.76


ratio 2.618:1. This confirms the theoretical implications of the goldenspiral that we explored in Chapter 10.

The 2.618 formula can be calculated either in terms of simpleprice changes or in terms of percentage changes. The situationdepends on the market being tracked and circumstances. The point,however, is that the formula provides a phenomenally accurateforecast of potential turning points.

NOTES

1. This is mathematically equivalent to saying that each diameter is related to theone that precedes it at 180º (that is, the next smallest diameter along the samestraight line) by the ratio 2.618.

2. Note that the first wave is excluded from the calculations because it does notinvolve a full diameter.

3. The number 34 is a Fibonacci number.4. Treasury 2½ per cent (redeemable on or after 1975) is called an ‘irredeemable’

on the assumption that yields will ‘never’ fall to 2½ per cent again. Only at thisstage would the stock actually be redeemed. In the interim, it constitutes cheapfunding.

5. The pull to redemption is the tendency of a stock to gravitate towards itsredemption price (100.00) as it nears the maturity date. This obviously doesnot reflect what is happening to the whole market.

6. The calculations are all conducted using closing prices. The results are usuallyeven more accurate using intra-day highs and lows.

153

12

The shape of things tocome

INTRODUCTION

In the analysis of the last chapter, a definite price pattern emerged. Inbull markets, there was a base pattern followed by a rapidly risingimpulse wave, while in bear markets the pattern was reversed. Insofaras the patterns were deduced from bond markets, we can also inferthat the pattern applies both to prices and to yields. In other words, thepattern is symmetrical. There is thus a basic three-wave pattern thatuses, or is controlled by, the ratio 2.618:1. We shall now look at therelevant patterns in more detail.

EXAMPLES

Figure 12.1 shows the monthly price pattern in UK gilts that startedoff the great disinflationary bull market way back in 1975. The bondused is an ultra-long-dated low-coupon stock, Treasury 2½ per centwith a maturity date on or after 1 April 1975. It is known as an ‘irre-deemable’ because it is unlikely to be redeemed until redemptionyields fall below 2½ per cent. Otherwise it remains cheap funding forthe government. The stock is a useful one to use because its longmaturity ensures that its price patterns are not distorted by the pull toredemption. This influence normally causes the price of a stock tomove towards a price of 100 as it approaches final maturity.

The pattern shown in the chart consists of a base formation and animpulse wave. The primary waves are denoted by the white numberswith the dark background – namely 1, 2 and 3. However, it is relevant,and important, that the base formation in the pattern consists of sixwaves. These are denoted by the unblocked lettering, 1–2–3 up andA–B–C down. Finally, even the base formation within this pattern (ie,waves 1 and 2) consists of six waves. These are denoted by the Romannumerals and lower case letters, i–ii–iii up and a–b–c down. We thushave a nested set of patterns that completely confirm the more theo-retical analysis of earlier chapters.

This example is by no means unique. Figure 12.2 shows the price ofTreasury 2½ per cent between February 1983 and February 1994. Thiswas the long pattern that ran into the bond ‘crash’ of 1994. Again, abase formation (ie, 1–2) precedes an impulse wave (ie, 3). And, again,the base formation consists of six waves and the base formation withinthis base formation consists of six waves.

Just to show how the wave patterns nest within one another, Figure12.3 compares the up-phase of the 1983–90 base pattern in Figure 12.2with the whole pattern of Figure 12.2. That is, the period February1983 to April 1986 is compared with the whole period February 1983to August 1993. Obviously the time elapse of each of the patterns isdifferent. However, they can be equalized – or made comparable – byseparating the time axes and making them geometrically equivalent.Hence the shorter period (1983–1986) is shown on the lower hori-zontal axis and the longer period (1983–1993) is shown on the upper


14

15

16

17

18

19

20

21

22

23

24

De

c-7

4

Fe

b-7

5

Ap

r-7

5

Ju

n-7

5

Au

g-7

5

Oct-

75

De

c-7

5

Fe

b-7

6

Ap

r-7

6

Ju

n-7

6

Au

g-7

6

Oct-

76

De

c-7

6

Fe

b-7

7

Ap

r-7

7

Ju

n-7

7

Au

g-7

7

Oct-

77

De

c-7

7

Pric

e

UK Treasury 2½% 'Irredeemable'(Monthly close)

Source: Datastream

1

C

B

A

3

2

1

2

3

i

c

b

a

iii

ii

Fe

b-7

8

Figure 12.1 The UK gilt-edged market, December 1974 to February 1978

horizontal axis. In order to show how well the two patterns mirror eachother, both time periods are extended to catch the initial drop after theirrespective wave 3 highs.

The first thing to notice is the way that the three-up/three-down basepatterns of both formations, and the subsequent up-wave, coincide.The second thing to notice, therefore, is that the timings of the turningpoints within the patterns display a constant relationship to the whole

The shape of things to come 155

18

20

22

24

26

28

30

32

34

36

38

Fe

b-8

3

Au

g-8

3

Fe

b-8

4

Au

g-8

4

Fe

b-8

5

Au

g-8

5

Fe

b-8

6

Au

g-8

6

Fe

b-8

7

Au

g-8

7

Fe

b-8

8

Au

g-8

8

Fe

b-8

9

Au

g-8

9

Fe

b-9

0

Au

g-9

0

Fe

b-9

1

Au

g-9

1

Fe

b-9

2

Au

g-9

2

Fe

b-9

3

Au

g-9

3

Pric

eUK Treasury 2½% 'Irredeemable'

(Monthly close)Source: Datastream

1

C

B

A

3

2

1

2

3

i

c

b

a

iii

ii

Fe

b-9

4

Figure 12.2 The UK gilt-edged market, February 1983 to February 1994

22.5

23.5

24.5

25.5

26.5

27.5

28.5

29.5

30.5

Fe

b-8

3

Ap

r-8

3

Ju

n-8

3

Au

g-8

3

Oct-

83

De

c-8

3

Fe

b-8

4

Ap

r-8

4

Ju

n-8

4

Au

g-8

4

Oct-

84

De

c-8

4

Fe

b-8

5

Ap

r-8

5

Ju

n-8

5

Au

g-8

5

Oct-

85

De

c-8

5

Fe

b-8

6

Ap

r-8

6

Ju

n-8

6

Au

g-8

6

Oct-

86

Pric

e

20

22

24

26

28

30

32

34

36

38

40

Fe

b-8

3

Au

g-8

3

Fe

b-8

4

Au

g-8

4

Fe

b-8

5

Au

g-8

5

Fe

b-8

6

Au

g-8

6

Fe

b-8

7

Au

g-8

7

Fe

b-8

8

Au

g-8

8

Fe

b-8

9

Au

g-8

9

Fe

b-9

0

Au

g-9

0

Fe

b-9

1

Au

g-9

1

Fe

b-9

2

Au

g-9

2

Fe

b-9

3

Au

g-9

3

Fe

b-9

4

Au

g-9

4

Fe

b-9

5

Au

g-9

5

Pric

e

UK Treasury 2½% 'Irredeemable'(Monthly close)

Source: Datastream

1983-1986(l.h.scale, lower time axis)

1983-1995(r.h.scale, upper time axis)

Figure 12.3 Pattern comparisons in the gilt market, 1983 to 1986 and1983 to 1995

pattern of which they are a part. The patterns are therefore repro-ducing themselves.

PRICE AND MOMENTUM

One of the difficulties with this sort of analysis is that the patterns areusually very clear when a market is oscillating sideways, or is notbeing subjected to economic shocks. However, as soon as a sustainedtrend emerges, or a significant information shock appears, the shorter-term pattern tends to become obscured. Hence, for example, while ithas been possible to identify the 1–2–3–A–B–C pattern in the UK gilt-edged market during the price oscillations of 1975–1985, it becomesincreasingly difficult to do so once the effects of disinflation started toset in. In other words, we need a method of de-trending the data.

Historically, a great deal of attention has been paid to this aspect ofanalysis and different analysts will have different ideas on the subject.However, by far and away the simplest solution is to use a straight-forward momentum indicator, such as a percentage rate of change.A percentage rate of change has the dual advantage of being inde-pendent of the level of the market and being easy to calculate. Undersome circumstances, an analyst may want to smooth the data by usingsome form of moving average before applying the momentum calcu-lations. However, this is by no means essential.

The advantages of using a momentum indicator can be demon-strated with reference to the US Treasury bond market. Shown inFigure 12.4 is a weekly price index of constant maturity 10-year T-bonds between May 1988 and April 1990. The index is calculatedby the simple expedient of deducting the yield (which is availablefrom the US Federal Reserve Board) from 100. Using constantmaturity bond yields has the advantage – as with Treasury 2½ per centabove – of avoiding the problems of the pull to redemption. It canimmediately be seen that the price index pattern is very similar tothose generated by the UK gilt-edged market. Further, because thepattern applies to a relatively short and undistorted period of time, thearchetypal six-wave (1–2–3–A–B–C) pattern can very clearly bediscerned.

Now look what happens (see Figure 12.5) when a 13-weekpercentage rate of change is calculated from the index. The six-wavepattern actually becomes clearer. There is an issue about the fact thatthe momentum low in March 1990 precedes the price low; but thisphenomenon will be dealt with in Chapter 13.


MOMENTUM EXAMPLES

Having made the adjustment to momentum patterns, we can now lookat some additional charts of the US T-bond market. Hence, Figures12.6 to 12.9 show the 13-week percentage change in the price index of10-year constant maturity bonds for March 1990 to April 1992, April1992 to May 1994, May 1994 to May 1996 and May 1996 to May1998. In each case, the six-wave pattern (which, as before, is denoted1–2–3–A–B–C) is clearly visible.


90

90.5

91

91.5

92

92.5

May-8

8

Ju

n-8

8

Ju

l-8

8

Au

g-8

8

Se

p-8

8

Oct-

88

No

v-8

8

De

c-8

8

Ja

n-8

9

Fe

b-8

9

Ma

r-8

9

Ap

r-8

9

May-8

9

Ju

n-8

9

Ju

l-8

9

Au

g-8

9

Se

p-8

9

Oct-

89

No

v-8

9

De

c-8

9

Ja

n-9

0

Fe

b-9

0

Ma

r-9

0

Ap

r-9

0

Ind

ex

1

2

3

A

B

C

1

2C

B

A

3

US 10-Year Treasury Bonds(100-constant maturity yield, weekly, daily averages)

Source: Federal Reserve Board

90

90.5

91

91.5

92

92.5

May-8

8

Ju

n-8

8

Ju

l-8

8

Au

g-8

8

Se

p-8

8

Oct-

88

No

v-8

8

De

c-8

8

Ja

n-8

9

Fe

b-8

9

Ma

r-8

9

Ap

r-8

9

May-8

9

Ju

n-8

9

Ju

l-8

9

Au

g-8

9

Se

p-8

9

Oct-

89

No

v-8

9

De

c-8

9

Ja

n-9

0

Fe

b-9

0

Ma

r-9

0

Ap

r-9

0

Pric

e in

dex

-1.5

-1

-0.5

0

0.5

1

1.5

2

13-w

eek

% c

han

ge

1

2

3

A

B

C



Price index(l.h.scale, lower

time axis)

13 wk % change(r.h.scale, upper

time axis)

Figure 12.4 The US Treasury bond market, May 1988 to April 1990

Figure 12.5 T-bond price and momentum, May 1988 to April 1990


-1.5

-1

-0.5

0

0.5

1

1.5

2

Ma

r-9

0

Ap

r-9

0

May-9

0

Ju

n-9

0

Ju

l-9

0

Au

g-9

0

Se

p-9

0

Oct-

90

No

v-9

0

De

c-9

0

Ja

n-9

1

Fe

b-9

1

Ma

r-9

1

Ap

r-9

1

May-9

1

Ju

n-9

1

Ju

l-9

1

Au

g-9

1

Se

p-9

1

Oct-

91

No

v-9

1

De

c-9

1

Ja

n-9

2

Fe

b-9

2

Ma

r-9

2

13-w

eek

% c

han

ge



1

2

3

A

B

C

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

Apr-

92

May-9

2

Jun-9

2

Jul-92

Aug-9

2

Se

p-9

2

Oct-

92

No

v-9

2

De

c-9

2

Ja

n-9

3

Fe

b-9

3

Ma

r-9

3

Apr-

93

May-9

3

Jun-9

3

Jul-93

Aug-9

3

Se

p-9

3

Oct-

93

No

v-9

3

De

c-9

3

Ja

n-9

4

Fe

b-9

4

Ma

r-9

4

Apr-

94

May-9

4

13-w

eek

% c

han

ge



1

2

3

A

B

C

Figure 12.6 T-bond momentum pattern, March 1990 to April 1992

Figure 12.7 T-bond momentum pattern, April 1992 to May 1994

INTIMATIONS OF CYCLICALITY

It may already have been noticed that the sequence of charts showingmomentum in the T-bond market are contiguous. The five chartsreveal a six-wave pulse for a succession of (roughly) two-yearperiods. We therefore have some evidence that not only does each pairof up- and down-waves evolve according to an archetypal pattern but


-1.5

-1

-0.5

0

0.5

1

1.5

2

May-9

4

Ju

n-9

4

Ju

l-9

4

Au

g-9

4

Se

p-9

4

Oct-

94

No

v-9

4

De

c-9

4

Ja

n-9

5

Fe

b-9

5

Ma

r-9

5

Ap

r-9

5

May-9

5

Ju

n-9

5

Ju

l-9

5

Au

g-9

5

Se

p-9

5

Oct-

95

No

v-9

5

De

c-9

5

Ja

n-9

6

Fe

b-9

6

Ma

r-9

6

Ap

r-9

6

13-w

eek

% c

han

ge



1

2

3

A

B

C

-1.5

-1

-0.5

0

0.5

1

1.5

2

May-9

6

Ju

n-9

6

Ju

l-9

6

Au

g-9

6

Se

p-9

6

Oct-

96

No

v-9

6

De

c-9

6

Ja

n-9

7

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

May-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

Oct-

97

No

v-9

7

De

c-9

7

Ja

n-9

8

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

13-w

eek

% c

han

ge



1

2

3

A

B

C

Figure 12.8 T-bond momentum pattern, May 1994 to May 1996

Figure 12.9 T-bond momentum pattern, May 1996 to May 1998

that the period between successive lows is not random. In other words,the T-bond market appears to be driven by structured rhythms.

We need to leave a more detailed analysis of this until later,because we are here dealing specifically with patterns. However, wecan already see that the theoretical framework of the price pulse thatwas explored in earlier chapters has some grounding in the evidence.Price fluctuations contain a six-wave pattern and are subject toregular periodicities.

STYLIZED PATTERNS

The six-wave pattern that we have so far analysed is very simple. Ingeneral, a recovery from a low is partially retraced. This is thenfollowed by an impulse wave. Eventually, prices fail to make anyfurther headway and drop. This drop is partially retraced. Then there isa dynamic impulse wave downwards.

It follows, however, from what has previously been said, and thendemonstrated in the examples from the UK gilt market, that lower-level oscillations will basically reflect the patterns of higher-leveloscillations. Hence the base and top patterns will themselves consistof six waves and will thus be a smaller version of the larger pattern ofwhich they are a part. This is a good example of ‘fractal geometry’.1 Astylized version of the net result is shown in Figure 12.10.

ASYMMETRY

The unfolding of lower-level patterns is continuously being modifiedby the influence of higher-level patterns. This results, from time totime, in clear distortions. These distortions can be particularlyprofound in bear markets, where top formations may sometimes besuppressed and where the downward-flowing impulse wave may bevery short in time but long in price. This is due to the asymmetric


Figure 12.10 Stylized T-bond price cycle

influence of fear. Specifically, the fear of losing money will oftenencourage long base patterns and create a slowly rising impulse wave.However, it will also encourage quicker reactions when things start togo wrong.

CONCLUSION

It is apparent from this analysis that the actual shape of the price pulsewill be distorted by higher-order trends. These distortions give anillusion of randomness to what is actually an ordered process.

The archetypal pattern is a three-phase movement, whether up ordown. The first two phases constitute the top pattern or the basepattern. The third phase consists of a dynamic impulse wave. What ismore, the pattern repeats itself as the market evolves through time.This pattern is central to the phenomenon of cycles.

NOTE

1. Fractal geometry, which reflects the idea that lower-degree details of a naturalphenomenon are identical to higher-degree details of the same phenomenon,receives a great deal of support from the scientific community. The idea offractal geometry was pioneered by Professor Mandelbrot at the IBM ResearchLaboratories. It suggests that, no matter how much an object is magnified, thenew details appearing on the smaller scale are always the same as the olddetails on the larger scale. See Mandelbrot, Benoit (1982) The FractalGeometry of Nature, W H Freeman, New York.


163

Part Three

Forecasting turningpoints

164


165

13

The phenomenon ofcycles

INTRODUCTION

One of the most difficult questions facing a technical analyst is how todifferentiate a signal that requires a minor portfolio adjustment from asignal that necessitates a major re-appraisal of a strategic view. Thetraditional answer has been that it doesn’t matter insofar as accurateshort-term decisions will ensure that the long-term decisions arelargely unnecessary. This answer is probably correct for day-tradersand small-scale investors. However, reliance on the short-term islargely incorrect for strategic investors and large-scale moneymanagers: the movement of large sums of money can incur penaltransaction costs, is often disallowed under management agreementsand inevitably exposes funds to huge performance risks againstcompetitors. Long-term managers have to take long-term decisions,so even the most enthusiastic of technical fund managers will castmore than a passing glance at so-called economic ‘fundamentals’.

This – and the fact that technical analysis still does not have a satis-factory theoretical underpinning amongst academics – helps toexplain why economic analysis continues to have such an intractablehold over investment decisions. Of course, it wouldn’t matter ifeconomic forecasting was accurate but – as we argued in Chapter 6 –it isn’t. First, the intellectual framework of economics is still largelyincapable of placing accurate time frames on turning points in

relevant variables. Second, economic forecasting is based on data thatare both out of date and inaccurate. Indeed, at the point where majorinvestment decisions are likely to be made, economic analysis oftenprovides no more than an intellectual comfort for an emotionally basedguess.

So we’re left with a situation where short-term technical analysismay not be clear about the importance of a signal and where long-termeconomic forecasting may not be clear about the imminence of aturning point. Is there any way that the resulting uncertainty can bereduced? The answer is: yes. And it comes from the realm of timecycle analysis.

THE INFLUENCE OF GROUPS

In the earlier chapters of this book, it was argued that the conflictbetween random fluctuations and ordered process in markets is essen-tially resolved by the influence of the group. That is, the inherentdesire of otherwise independent individuals to participate in greaterwholes allows the ‘whole’ to devolve order onto the participants. Theinfluence of the whole is psychological, emotional and, accordingly,powerful: individuals are eventually induced to do what everyone elseis doing. This is why large numbers of people simultaneously buydotcom stocks or second houses to rent, when reason suggests thatcaution might be more appropriate.

So, if we can allow that each person has a tendency both to be anindividual and to be part of a group, then two of the fundamentalassumptions of economic theory fall flat on their face. First, aggregatebehaviour is not just the sum of the parts; it is something significantlymore. Second, this augmented behaviour involves an important non-rational (and sometimes even irrational) dimension. In effect, whateconomic theory ignores is that each of us has an impulse to create‘meaning’ for ourselves, where meaning is measured by our innerworld of feelings, and where feelings are stimulated by, and integratedwith, our relationships with others.

Of course, it can be argued that the simplifying assumption ofrational behaviour by independent individuals just makes the fore-casting models easier to use. The problem, however, is that it alsoruns the risk of making the forecasts less accurate. Critically, thesimplifying assumption that behaviour is rational and non-affectivedoes not allow for the simple reality of human behaviour that lies atthe heart of cyclical behaviour – the emergence of satiation. Indeed,

166 Forecasting turning points

in economics, satiation is treated as an aberration: the unspokenpresumption is that we never have enough ‘goods’ (note the termi-nology). Hence, growth can – and should – continue persistently andindefinitely, subject only to the birth rate (which creates new entrantsto the consumer market) and the ability of the system to generate new‘goods’ (through innovation).

SATIATION

In nature, however, every living system oscillates between a state ofactivity and a state of rest. Indeed, rest is essential in order to re-energize an organism and delay entropy. Without rest, life (as we knowit) could not exist. So, cycles are the signature of life. The question is,how does an organism ‘know’ when to switch from activity to rest, orback again? As indicated in Chapter 9, the answer is satiation:‘enough’ of activity causes an organism to switch from activity intorest, while ‘enough’ of rest causes an organism to switch from rest intoactivity. As we approach satiation at the end of a day, we begin to loseenergy, our body heat drops and our metabolism slows. Then,suddenly, we fall asleep. We traverse a process gap. Conversely, whenwe’ve had enough sleep, our energy potential is re-established, so ourorgan systems start to speed up and our body temperature rises. Again,we experience a process gap; and, suddenly, we are awake. Everyliving organism regularly, and at some infinitesimal moment in time,passes over the cusp that delineates sleep from rest. Moreover, thesemoments in time are entrained across nature. Most of us, for example,go to sleep before midnight and wake up in the early morning (give ortake a late night out!). The harmonization of activity and restthroughout the biosphere reduces interference and allows total energyusage to be minimized.

TRANSFORMATION OF ENERGY ANDINFORMATION

There is, however, another element here, which is also critical to ourunderstanding of oscillations in living systems. Living systemsrespond to the input of energy and information. The interesting point,however, is that when a system receives new energy or information, ithas to divert existing energy to deal with it. A simple example is eating.

The phenomenon of cycles 167

Food provides new energy to our bodies. The first thing that happens,however, is that energy resources are diverted to deal with the transfor-mation and absorption of the food. That is, our freely available energytends to drop and, while this is happening, we actually tend to slowdown and feel sleepy. In a sense, this slowdown revisits the drop inenergy that initially signalled the need to eat. But there is a huge qual-itative difference: the slowdown without food marks the end of aphase, while the slowdown with food marks the transition to a newphase. Having eaten, our energy is eventually considerably enhancedand we can go about our activities with some degree of vigour.

Obviously what we have here is a three-phase process that proceedsthrough input, absorption and application. Importantly, this patterndoes not just apply to energy; it also applies to information. Asdiscussed in Chapter 9, the diversion of energy to deal with the transferof information from short-term memory to long-term memorymomentarily interrupts the ability to apply new learnings. So, thethree-wave pattern is also the signature of ‘learning’. The first twostages are the true learning stages (where information actually altersthe qualitative structure of the organism) and the third stage constitutesthe confirmation that learning has occurred.

There are thus four ideas that can be used to produce a model ofoscillations that mirrors what happens in financial markets. Theseideas are: 1) individuals combine into groups; 2) groups respond tothe input of energy and information from their environment; 3) theassociated absorption of new energy and information involves apause in activity; and 4) satiation causes a polarity switch betweenactivity and rest. One inference, of course, is that a financial marketcrowd or an economically active group can be treated as if they wereliving organisms.

THE PATTERN THAT CONNECTS

These four ideas are the basis of the price pulse, with its six-wave(1–2–3–A–B–C) pattern. The important implication, therefore, is thata full theory of cyclical behaviour needs to take into account the factthat rhythmic oscillations have both a definite mechanism and aprecise pattern. Historically, great effort has been spent in identifyingtime windows and precise dates when a cycle inflexion might beexpected to occur. Great additional effort has then been spent in tryingto identify the internal and external influences that might validateprecise periodicities. Almost inevitably, however, the efforts have only


been partially successful and, as a consequence, analysts have tendedeither to give the influence of time a very low priority or to ignore italtogether. But, as already argued, fluctuations are essential to theexistence of life.

A potential solution, therefore, to the problem of variable periodic-ities is to track directly the pattern of the oscillations through time. Ifthe pattern reflects (or is consistent with) the influence of the sati-ation-and-energy gap mechanism, then there is a good chance that wecan know where we are in the cycle. The average historical periodicityof the market’s heartbeat can be used to estimate when an inflexionpoint might arise. Ultimately, however, it is the pattern that determinesthe imminence of, and confirms the passing of, a turning point.

We shall be developing this theme in far more detail in the chaptersahead. In the meantime, it is worth noting that the existence of anarchetypal pattern in cycles can be used to cross-correlate cycles withone another. Short-term cycles can be compared with long-term ones,adjacent cycles can be compared with one another, and cycles fromdifferent markets can be compared with one another. The point is that,as a new cycle is evolving, it can be compared directly and (more orless unambiguously) with appropriate cycles that have alreadycompleted. The price pulse is, to use Gregory Bateson’s phrase,1 thepattern that connects.

TRACKING THE CYCLE

The first consideration, therefore, is: how can the pattern of a cyclebest be tracked? In Chapter 12, we had a brief look at the pattern of theprice pulse in the context of bond markets. It was quite clear from theanalysis that the pattern can be deduced directly from the behaviour ofprices (or yields). The problem, however, is that, when a market isstrongly trending, then the pattern may be difficult to see. Thesolution, obviously, is to de-trend the data. Traditionally, the preferredmethod for doing this is to use deviations from some form of movingaverage. The simplest method, however, is to use a momentum index.

MOMENTUM INDICES

A momentum indicator measures the rate of change in prices per unitof time. It therefore measures the speed of a market. The nature and


time period of the indicator that is used are, to some extent, amatter of personal preference. However, there are two simpleguidelines that can be used. First, it is better to use a percentagerate of change if there has been a large change in levels of theunderlying index. A 10-point move in an index means somethingdifferent if the base is 25 points rather than 100 points. Otherwisea simple difference can be used, calculated by subtracting one termin the data series from another.

Second, the time period that is used should be as short as isconvenient. Long time periods have a tendency to iron out short-term fluctuations and short time periods have a tendency toemphasize them. However, the period should not be more than 25per cent of the time period being tracked. Hence for a four-yearcycle (for example), a one-year rate of change indicator is about thelongest that should be used.

MOMENTUM AND THE CYCLE

One of the strengths of a momentum indicator is that it can provide agood indication of the actual status of a cycle in ‘real’ time. There is avery good theoretical reason for this. Figure 13.1a shows the positionof a regular cycle with respect to time. Note that the slope of the cyclevaries with the stage of the cycle. At point A, the slope is flat becausethe cycle is in the process of turning down. At point B, the slope issteep and downward (or negative). At point C, the slope is flat againbecause the cycle is turning up. And at D, the slope is steep andupward (or positive).

Now, the steepness of the curve represents the rate at which theposition changes. It is the velocity of the cycle and is known to math-ematicians as the first derivative of the cycle with respect to time. Thevelocity is shown in Figure 13.1b. The important point to note is thatthe resulting curve is the same shape as the original cycle but it isdisplaced a quarter-cycle back from the original.

The analysis can be taken a stage further. The slope of the velocitycycle represents the rate at which velocity is changing. It is thereforethe acceleration/deceleration of the cycle and is known to mathemati-cians as the second derivative of the cycle with respect to time. Theacceleration/deceleration cycle is shown in Figure 13.1c. Note thatthis cycle is displaced a quarter-cycle ahead of the velocity cycle. It istherefore a half-cycle back from the original cycle.



Figure 13.1 Change, velocity and acceleration

ANTICIPATING INFLEXION POINTS

This is not, of course, the whole story, because price cycles do not takethe smooth curvilinear form shown in Figure 13.1. Nevertheless, itdoes give us access to some important information. First, a ‘spike’ inmomentum will occur at least a quarter-cycle ahead of the cycle beingtracked. This can give us an estimate for the timing of the absoluteturning point. If we are tracking a 3.25-year cycle and an ‘oversold’reading occurs in the momentum indicator, we can estimate that theabsolute low might occur approximately 42 weeks (a quarter of the3.25-year cycle) later. As a corollary, the time elapse between the pointwhen a market becomes ‘oversold’ (or ‘overbought’) and the pointwhere a reversal in the absolute price level occurs can be used toestimate a possible periodicity for a new cycle. All that needs to bedone is to multiply the time elapse by four. Finally, a distinct warningabout the possible presence of speculative – or, at least, highlyemotional – activity in the market can be given if prices continue totrend well beyond the time elapse that normally follows a momentumreversal.

Hence, in Figure 13.2, the Dow momentum reversals in January1983 and August 1989 were followed 10 or 11 months later by priceinflexions on weaker momentum. This was consistent with the oper-ation of an underlying 40- to 44-month cycle. In between these twodates, there was another momentum reversal followed by a priceinflexion. However, the quality of the relationship was very different.The absolute momentum peak occurred in March 1986. There wasthen a lower momentum peak 12 months later. But it was not accom-panied by a price peak. This could have been taken as indicating that aproblem was developing. The 12-month time elapse between themomentum peaks was actually consistent with a 48-month cyclerather than a 40- to 44-month cycle, and was therefore suggestingsome distortion. In the event, prices continued to rise into a high inAugust 1987, which was 17 months after the initial momentuminflexion. What followed was the 1987 ‘crash’.

VELOCITY AND NON-CONFIRMATION

Following on from this, the relationship between prices and velocityhelps to explain the idea of ‘non-confirmation’ that is so important tomarket analysis. It is apparent from Figure 13.2 that upward


momentum tends to be weakening at a price high and that downwardmomentum tends to be fading at a price low. In fact, for longercycles, the pattern is likely to be quite clear cut. There will be apreliminary spike in short-term momentum as the market becomesoverextended. This will be followed by a sharp correction in themarket. Prices then move back towards the level at which theybecame overextended. As prices move out into new territory, therewill be a second spike in the momentum indicator, which will notusually be greater than the first one.

Consequently, the second higher peak or lower trough in prices is‘not confirmed’ by the associated momentum indicator. Under normalcircumstances, this warns that the reversal is important.

ACCELERATION AND THE CYCLE

Finally, the acceleration/deceleration cycle provides us with a poten-tially powerful tool for confirming when a cycle has actually reversed.There are occasions when momentum stretches out to exceptionallevels (positive or negative) immediately after a reversal in absoluteprices. As we have seen, the cycle of acceleration occurs in oppositionto the cycle in absolute prices and a surge in momentum is thereforeprima-facie evidence that a high-level cycle has reversed direction.


-35

-30

-25

-20

-15

-10

-5

0

5

10

15

20

25

30

35

40

45

Se

p-8

1

Ja

n-8

2

May-8

2

Se

p-8

2

Ja

n-8

3

May-8

3

Se

p-8

3

Ja

n-8

4

May-8

4

Se

p-8

4

Ja

n-8

5

May-8

5

Se

p-8

5

Ja

n-8

6

May-8

6

Se

p-8

6

Ja

n-8

7

May-8

7

Se

p-8

7

Ja

n-8

8

May-8

8

Se

p-8

8

Ja

n-8

9

May-8

9

Se

p-8

9

Ja

n-9

0

May-9

0

Se

p-9

0

6-m

on

th %

ch

ang

e

600

1100

1600

2100

2600

3100

DJI

A in

dex

Momentum(l.h.scale)

Dow Index(r.h.scale)

Data source: Bloomberg

Figure 13.2 Momentum and price in the Dow Jones Industrial Average,1981 to 1990

It is important, however, to be clear that the acceleration inmomentum is not just a result of mathematics – for example, themomentum index was previously ‘oversold’ and is now ‘overbought’.Sometimes, therefore, there will be some doubt. However, if theacceleration in momentum occurs after a non-confirmed priceextreme, then the chances are that the signal is genuine. An example isshown in Figure 13.3. Here the low in the Dow in late 1975 was notconfirmed by a two-month rate of change indicator. The subsequentsurge in that rate of change indicator was strong evidence that a turnhad occurred.

CONCLUSION

This chapter has recapitulated the influence of the three-wave patternthat regulates the processes of learning and energy absorption in theevolution of price–time cycles. The basic pattern applies both tophases of expansion and phases of contraction. Importantly, behav-ioural cycles switch polarity between expansion and contraction whenthey transit into an energy gap. This simple mechanism – which,unfortunately, is usually ignored simply because it is so simple –applies to all cycles. This, in turn, means that it is the pattern of thecycle that is the critical element in timing the turn, not the actual


-30

-20

-10

0

10

20

30

40

50

Ja

n-7

4

Fe

b-7

4

Ma

r-7

4

Ap

r-7

4

May-7

4

Ju

n-7

4

Ju

l-7

4

Au

g-7

4

Se

p-7

4

Oct-

74

No

v-7

4

De

c-7

4

Ja

n-7

5

Fe

b-7

5

Ma

r-7

5

Ap

r-7

5

May-7

5

Ju

n-7

5

Ju

l-7

5

Au

g-7

5

Se

p-7

5

Oct-

75

No

v-7

5

2 m

th %

ch

ang

e

0

100

200

300

400

500

600

700

800

900

1000

DJI

A in

dex2 mth % change

(l.h.scale)

DJIA index(r.h.scale)

Data source: Datastream

Figure 13.3 Non-confirmation and accelerating momentum in the Dow,1974 to 1975

passage of time itself. The latter is important insofar as previous beatsof a cycle have established a periodicity of some sort. This will enablethe analyst to calculate time windows in which a reversal can beexpected. However, the periodicity will be variable and, ultimately,the periodicity is subordinate to the pattern.

The pattern of the cycle can be tracked in real time, using a simplemomentum indicator. Once the three-up/three-down pattern has beenidentified, it should be possible to anticipate the pattern of an evolvingcycle beat, both directly and by comparing it with historicalprecedent. We shall now look at this process in more detail.

NOTE

1. Bateson, Gregory (1979) Mind and Nature: An essential unity, WildwoodHouse, London.


176

14

The threefold nature ofcycles

INTRODUCTION

We can now look at some of the evidence for the existence of specificcycles in financial markets. This will also give us the opportunity toexplore the basic characteristics of such cycles and to draw conclusionsthat can be used for wider analyses. To start, therefore, we shall concen-trate on the US Dow Jones Industrial Average. There are very strongreasons for doing this, not the least of which is that it is the most well-known index within the context of history. Modern markets, with theirwidespread use of futures and options, may be more directly orientatedto other indices. However, it was the Dow that was followed during theWall Street Crash of 1929; it was the Dow that hit the headlines in theCrash of 1987; and it is the Dow that is treated as the pre-eminent indexon the TV news channels and in the daily press. The point is that theDow Jones Industrial Average is the one that has reflected investoremotions in the equity market over a long span of history. If there is arhythm in the US stock market, it should exist in the Dow.1

THE 11-YEAR CYCLE IN THE DOW

There is a great deal of evidence, which is borne out by the detailedwork of many others,2 that one of the dominant cycles in the Dow

The threefold nature of cycles 177

Jones Industrial Average has an average duration of about 11 years.The cycle seems to have undergone a big distortion during the SecondWorld War – which is not surprising – so there is some doubt aboutwhether the correct timing of the low that was due towards the end ofthe war actually occurred in the autumn of 1943 or in the autumn of1946. Despite this, subsequent lows occurred in winter 1957, summer1970, autumn 1981, autumn 1990 and autumn 2001.

The 11-year cycle in the Dow can be tracked using a six-monthpercentage rate of change in the monthly closes. Figures 14.1 to 14.5therefore show the resulting profile over successive periods of approxi-mately 11 years, starting in November 1943. Figure 14.1 covers the 158months from November 1943 to February 1957. Figure 14.2 covers the160 months from February 1957 to June 1970. Figure 14.3 coves the135 months from June 1970 to September 1981. Figure 14.4 covers the110 months from September 1981 to November 1990. And Figure 14.5covers the 131 months from November 1990 to October 2001.

In each of the figures, an initial attempt has been made to identifythe archetypal six-wave pattern that defines a complete beat of theprice pulse. The resulting conclusions will be subject to alteration oncloser consideration; but it is exactly the sort of approach that ananalyst might make when first confronted with the patterns. Hence, ineach figure, the archetypal pattern is distinguished by the1–2–3–A–B–C notation that has already been used. The onlyallowance that has been made at this stage is for the fact that, after

-25

-20

-15

-10

-5

0

5

10

15

20

25

30

35

40

Nov

-43

May

-44

Nov

-44

May

-45

Nov

-45

May

-46

Nov

-46

May

-47

Nov

-47

May

-48

Nov

-48

May

-49

Nov

-49

May

-50

Nov

-50

May

-51

Nov

-51

May

-52

Nov

-52

May

-53

Nov

-53

May

-54

Nov

-54

May

-55

Nov

-55

May

-56

Nov

-56

6 m

onth

per

cent

age

chan

ge

1

C

B

A

3

20

The Dow Jones Industrial Average(Monthly, 6-month % change)

Data source: Economagic

Figure 14.1 The Dow, November 1943 to February 1957

deep momentum lows at the end of a cycle, the initial upswing may beparticularly dynamic. This means that the top of this initial surge willnot necessarily mark the peak of the cycle. This consideration appliesto Figures 14.2 and 14.3.

In the last cycle – ie, 1990–2001 – the pattern is obviously morecomplicated than that of its predecessors. This is due to the fact thatthe pattern evolved during a particularly dynamic period of economicgrowth, where one of the underlying themes is trial and error.


-30

-20

-10

0

10

20

30

Feb-

57

Aug

-57

Feb-

58

Aug

-58

Feb-

59

Aug

-59

Feb-

60

Aug

-60

Feb-

61

Aug

-61

Feb-

62

Aug

-62

Feb-

63

Aug

-63

Feb-

64

Aug

-64

Feb-

65

Aug

-65

Feb-

66

Aug

-66

Feb-

67

Aug

-67

Feb-

68

Aug

-68

Feb-

69

Aug

-69

Feb-

70

6-m

onth

per

cent

age

chan

ge

1

C

B

A

3

20



-40

-30

-20

-10

0

10

20

30

40

50

Jun-

70

Oct

-70

Feb-

71

Jun-

71

Oct

-71

Feb-

72

Jun-

72

Oct

-72

Feb-

73

Jun-

73

Oct

-73

Feb-

74

Jun-

74

Oct

-74

Feb-

75

Jun-

75

Oct

-75

Feb-

76

Jun-

76

Oct

-76

Feb-

77

Jun-

77

Oct

-77

Feb-

78

Jun-

78

Oct

-78

Feb-

79

Jun-

79

Oct

-79

Feb-

80

Jun-

80

Oct

-80

Feb-

81

Jun-

81

6-m

onth

per

cent

age

chan

ge 1

C

B

A

3

20



Figure 14.3 The Dow, June 1970 to September 1981

Figure 14.2 The Dow, February 1957 to June 1970

Increased risk often means increased volatility because things can –and do – go wrong. We shall analyse the pattern in more detail at alater stage. In the meantime, the initial labelling of the 1990–2001pattern is as indicated in Figure 14.5.

IDEALIZED CYCLESA useful starting point for analysing these charts is the construction ofan ‘idealized’ timing pattern, which might apply to them all. Figure


-30

-20

-10

0

10

20

30

40

Sep

-81

Jan-

82

May

-82

Sep

-82

Jan-

83

May

-83

Sep

-83

Jan-

84

May

-84

Sep

-84

Jan-

85

May

-85

Sep

-85

Jan-

86

May

-86

Sep

-86

Jan-

87

May

-87

Sep

-87

Jan-

88

May

-88

Sep

-88

Jan-

89

May

-89

Sep

-89

Jan-

90

May

-90

Sep

-90

6-m

onth

per

cent

age

chan

ge

1

C

B

A

3

20



-15

-10

-5

0

5

10

15

20

25

30

Nov

-90

Mar

-91

Jul-9

1

Nov

-91

Mar

-92

Jul-9

2

Nov

-92

Mar

-93

Jul-9

3

Nov

-93

Mar

-94

Jul-9

4

Nov

-94

Mar

-95

Jul-9

5

Nov

-95

Mar

-96

Jul-9

6

Nov

-96

Mar

-97

Jul-9

7

Nov

-97

Mar

-98

Jul-9

8

Nov

-98

Mar

-99

Jul-9

9

Nov

-99

Mar

-00

Jul-0

0

Nov

-00

Mar

-01

Jul-0

1

6-m

onth

per

cent

age

chan

ge

1

C

B

A

3

20



Figure 14.4 The Dow, September 1981 to November 1990

Figure 14.5 The Dow, November 1990 to October 2001

14.6 shows such a pattern. The whole (higher-level) cycle is repre-sented by the move from 0 to C – ie, 1–2–3 up and A–B–C down. Itthen consists of three lower-level (sub-) cycles, each of which itselfcontains the archetypal six-wave pattern. In theory, each of theselower-level cycles will itself consist of three cycles. In other words,the cycles are nested within each other. In all cases, significant lowscan be expected to occur one-third and two-thirds along the timeelapse of the next higher cycle that contains it. Similarly, importanthighs occur at one-sixth, one-half and five-sixths along the timeelapse of that higher-level cycle.

COMPARING THE CYCLES

We can now compare each of the cycles shown in Figures 14.1 to 14.5to the idealized version shown in Figure 14.6. This is done in Figures14.7 to 14.11. In each figure, three things are held constant: thevertical scale that measures the six-month percentage rates of change;the vertical lines that represent the idealized timings of the peaks andtroughs relative to the total time elapse of each cycle; and the idealbehavioural pattern of each cycle, shown as the dashed line. In thisway, we have a blueprint on to which we can overlay each completeDow cycle.

Figure 14.7 shows the period from November 1943 to February1957. The first peak seems clearly defined inasmuch as it is followedby a large fall. However, the end of that fall probably does not mark


-35

-25

24 72 126

1

C

B

A

3

20

17% 50% 83%

100%67%33%0%

Figure 14.6 Idealized triadic cycle pattern

the proper end of the sub-cycle. The ideal behavioural pattern suggeststhat the actual low was in February 1946. Next, the sharp rally into thewave 3 peak seems to have begun on time but it probably ended a littleearly, in December 1949. Then there was the sequence of the initialdrop in momentum, the recovery into early 1951, and the downswingto the base of wave A. Obviously, there is some doubt about the exacttiming of the low of wave A because the ideal location is straddled bytwo momentum lows – one in early 1952 and one in mid-1953.However, the ideal shape of the subsequent rally into the peak of waveB suggests that the former is the correct one. Finally, momentum fallsinto early 1957. Hence, although this cycle was undoubtedly distortedby its placement in relation to the war so that we have had to makesome adjustments to our initial assessment of the timings within thecycle, the basic pattern was actually not too different to the blueprint.

Figure 14.8 shows the period from February 1957 to June 1970. Itstarts well enough in relation to the blueprint, but wave 2 is distortedby a big rally in early 1961, with the result that the effective low isdelayed until mid-1962. The subsequent rally is very sharp, as itshould be following a deep sell-off, so the cycle peak (on lowermomentum) is placed in early 1964. Then there is the fall into the baseof wave A, which – like the base of wave 2 – is delayed until late 1967.Finally, momentum has a three-wave rally into a high in late 1968,followed by a three-wave fall. The cycle thus follows the blueprint interms of pattern, but its highs and lows regularly occur later than theideal because of the upward distortion that occurred in early 1961.


Figure 14.7 The Dow, November 1943 to February 1957

-35

-25

-15

-5

5

15

25

35

45

Nov

-43

May

-44

Nov

-44

May

-45

Nov

-45

May

-46

Nov

-46

May

-47

Nov

-47

May

-48

Nov

-48

May

-49

Nov

-49

May

-50

Nov

-50

May

-51

Nov

-51

May

-52

Nov

-52

May

-53

Nov

-53

May

-54

Nov

-54

May

-55

Nov

-55

May

-56

Nov

-56

6-m

onth

% c

hang

e

-35

-25

-15

-5

5

15

25

35

45

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108 114 120 126 132 138 144 150 156

Months since low

6-m

onth

% c

hang

e

1

C

B

A

3

20

Figure 14.9 tracks the market’s performance from June 1970 to thebig market low in September 1981. The period started with a bigjump in momentum, so the correct placement of the first high iswhere it should be, on the lower momentum of late 1971. The baseof wave 2 was a little delayed, but the top of wave 3 arrived more orless on time. Again, the correct timing of the high is not after theinitial surge from the low, but on the subsequent lower momentumpeak. The low of wave A arrives on time, as does the peak of wave Band the low of wave C. The cycle thus accords very closely with theideal blueprint.


-35

-25

-15

-5

5

15

25

35

45

Jan-

57

Jul-5

7

Jan-

58

Jul-5

8

Jan-

59

Jul-5

9

Jan-

60

Jul-6

0

Jan-

61

Jul-6

1

Jan-

62

Jul-6

2

Jan-

63

Jul-6

3

Jan-

64

Jul-6

4

Jan-

65

Jul-6

5

Jan-

66

Jul-6

6

Jan-

67

Jul-6

7

Jan-

68

Jul-6

8

Jan-

69

Jul-6

9

Jan-

70

6-m

onth

% c

hang

e

-35

-25

-15

-5

5

15

25

35

45

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108 114 120 126 132 138 144 150 156

Months since low

6-m

onth

% c

hang

e

1

C

B

A

3

20

Figure 14.8 The Dow, February 1957 to June 1970

Figure 14.9 The Dow, June 1970 to September 1981

-35

-25

-15

-5

5

15

25

35

45

Jun-

70

Oct

-70

Feb-

71

Jun-

71

Oct

-71

Feb-

72

Jun-

72

Oct

-72

Feb-

73

Jun-

73

Oct

-73

Feb-

74

Jun-

74

Oct

-74

Feb-

75

Jun-

75

Oct

-75

Feb-

76

Jun-

76

Oct

-76

Feb-

77

Jun-

77

Oct

-77

Feb-

78

Jun-

78

Oct

-78

Feb-

79

Jun-

79

6-m

onth

% c

hang

e

-35

-25

-15

-5

5

15

25

35

45

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108 114 120 126 132

Months since low

6-m

onth

% c

hang

e

1

C

B

A

3

20

Figure 14.10 shows the momentum of the Dow from September 1981to November 1990. Very little needs to be said about this particularcycle. Like its predecessor, it follows the blueprint very closelyindeed. Even the sub-cycles seem to be in synchrony with the ideal.

Figure 14.11 tracks the momentum of the Dow from November1990 to the October 2001 low. Unlike the cycles of its two prede-cessors, this cycle suffers from what seem to be distortions. The firstsub-cycle follows the blueprint very closely; but the following sub-cycle is biased upwards to the right, and the last sub-cycle is


Figure 14.11 The Dow, November 1990 to October 2001

Figure 14.10 The Dow, September 1981 to November 1990

-35

-25

-15

-5

5

15

25

35

45

Sep

-81

Jan-

82

May

-82

Sep

-82

Jan-

83

May

-83

Sep

-83

Jan-

84

May

-84

Sep

-84

Jan-

85

May

-85

Sep

-85

Jan-

86

May

-86

Sep

-86

Jan-

87

May

-87

Sep

-87

Jan-

88

May

-88

Sep

-88

Jan-

89

May

-89

Sep

-89

Jan-

90

May

-90

Sep

-90

6-m

onth

% c

hang

e

-35

-25

-15

-5

5

15

25

35

45

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108

Months since low

6-m

onth

% c

hang

e

1

C

B

A

3

20

-35

-25

-15

-5

5

15

25

35

45

Nov

-90

Mar

-91

Jul-9

1

Nov

-91

Mar

-92

Jul-9

2

Nov

-92

Mar

-93

Jul-9

3

Nov

-93

Mar

-94

Jul-9

4

Nov

-94

Mar

-95

Jul-9

5

Nov

-95

Mar

-96

Jul-9

6

Nov

-96

Mar

-97

Jul-9

7

Nov

-97

Mar

-98

Jul-9

8

Nov

-98

Mar

-99

Jul-9

9

Nov

-99

Mar

-00

Jul-0

0

Nov

-00

Mar

-01

Jul-0

1

6-m

onth

% c

hang

e

-35

-25

-15

-5

5

15

25

35

45

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108 114 120 126

Months since low

6-m

onth

% c

hang

e

1

C

B

A

3

20

extremely volatile. In fact, the pattern is very specific. The wholepattern is related to major changes to the US (and, indeed, global)economic infrastructure. Such cycles are biased upwards to the rightand are terminated by a three-wave contraction. Moreover, in generalterms, the pattern of the whole cycle will be similar to the pattern ofthe middle cycle. Fortunately for analysts, this phenomenon does notjust arrive out of the blue. It is part of the natural processes ofevolution, which can be anticipated by paying attention to theeconomic cycle itself. We shall be looking at this in Chapters 15 and16. In the meantime, we can anticipate the conclusions by saying thatthe correct low for the second sub-cycle in the 1990–2001 triad is inAugust 1998 and that the correct high for the third sub-cycle is inApril 1999.

ACTUAL CYCLE TIMINGS

We can now conduct an analysis of actual cycle timings, based on theabove observations. Table 14.1 shows the time elapses, measured inmonths, between the important turning points in Figures 14.7 to 14.11.It indicates that the average periodicity of the five cycles starting inNovember 1943 is 11.58 years. However, it is also clear that the peri-odicity has been contracting. The average of the last four cycles was11.17 years, and the average of the last three was only 10.4 years. Themost recent complete cycle took 10.92 years. So there is some vari-ability in the time elapse of the 11-year cycle.

Nevertheless, Table 14.1 reveals something very interesting.Despite the variability in the overall cycle lengths, there seems to be atendency for the peaks and troughs generated by the sub-cycles tomeet the timings indicated by the blueprint shown in Figure 14.6. Themain peaks and troughs occur on average at positions that are 15 percent, 35 per cent, 51 per cent, 67 per cent and 83 per cent of the totaltime elapse of the cycle. These are very close to the timings suggestedby the blueprint, which correspond to the percentages that arise ifunity is divided into six parts. That is, one-sixth is 17 per cent, two-sixths (or a third) is 33 per cent, three-sixths (or a half) is 50 per cent,four-sixths (or two-thirds) is 67 per cent and five-sixths is 83 per cent.

In other words, peaks and troughs in a cycle tend to conform to avery simple formula: major troughs occur at around 33 per cent and 67per cent of the total time elapse of a cycle, and peaks occur at around17 per cent, 50 per cent and 83 per cent of the cycle.


THE FOUR-YEAR CYCLE IN THE DOW

This analysis is consistent with the idea that cycles are groupedtogether in threes, or in balanced triads. Specifically, the main cycleconsists of three lower-level cycles, each of which has a duration thatis about one-third of the duration of the main one. We can confirm thisconclusion with respect to the 11-year cycle. The time elapse,measured in months, of each of the sub-cycles in Table 14.1 is shownin Table 14.2. The average of the first sub-cycles in each triad is 48months, the average of all the second cycles is 45 months, and theaverage of all third cycles is also 45 months. The average of all 15cycles is therefore 46 months.3 This is what is more popularly knownas the four-year cycle in the Dow.

TRIADS, DYADS AND ENERGY GAPS

One of the conclusions that emerges from this analysis is that thepattern of the cycle is at least as important as the actual time elapse ofthe cycle. Another conclusion is that the mid-point of the second cyclewithin any triad will coincide with the mid-point of the relevant over-arching higher-level cycle. This latter is important because it means


DatesMonths from start of cycle

1st high 1

1st low 2

2nd high 3

2nd low A

3rd high B

3rd low C

Nov 43 – Feb 57 26 (16%)

51 (32%)

73 (46%)

99 (63%)

135 (85%)

158 (100%)

Feb 57 – Jun 70 22 (14%)

64 (40%)

83 (52%)

114 (71%)

139 (87%)

160 (100%)

Jun 70 – Sep 81 23 (17%)

51 (38%)

69 (51%)

92 (68%)

110 (81%)

135 (100%)

Sep 81 – Nov 90 16 (15%)

32 (30%)

54 (50%)

76 (70%)

95 (88%)

110 (100%)

Nov 90 – Oct 01 18 (14%)

44 (34%)

74 (56%)

86 (72%)

101 (90%)

131 (100%)

Averages 21 (15%)

48 (35%)

71 (51%)

93 (67%)

116 (83%)

139 11.58 yrs

Table 14.1 Cycle timings in the Dow

that cycles are genuinely harmonically related to one another. If cycleswere instead grouped in pairs, there would be a conflict between theupper- and lower-level cycles. Specifically, the peak of the formerwould coincide with a low in the latter. In terms of integrated systems,this is simply a non-starter: it would indicate dissonance.Nevertheless, there must be one or more reasons why some analystsinsist that cycles arise in pairs. This needs to be addressed.

The first reason why cycles are often interpreted as being dyads isthat dissonance is exactly what does seem to appear. That is, cyclelows arise at approximately the point where a high is expected. Thisfrustrating phenomenon is taken by some to ‘confirm’ the presence ofdyads and is taken by others as being an indication that cycles simplydon’t work. There is, however, the possibility that the phenomenon isan illusion, arising from a misinterpretation of the evidence. Whatmay be happening is that the influence of an energy gap after a high isbeing taken as indicating the presence of a low.

Figure 14.12 demonstrates the essence of this problem. It focuseson momentum immediately after an important high. The upper chartassumes a higher-level cycle that is divided into three lower-levelcycles. At the peak of the second cycle the market is satiated. Ittherefore falls into an energy gap, which reverses the polarity of themarket into bearish mode. Prices accordingly fall sharply. In the lowerchart, this fall in prices registers as a sharp drop in momentum. Whathappens, therefore, is that the momentum indicator registers a low. So,if the analyst is measuring cycles from momentum low to momentumlow, then the data register the end of a cycle. Consequently, the wholehigher-level cycle (when it finally ends) looks as if it is constructed oftwo sub-cycles. In fact, what has happened is that the signs of a highhave been misinterpreted as a low. Ex post, the mid-cycle momentum


DatesLength (months)

Total 1st cycle 2nd cycle 3rd cycle Averages

Nov 43 – Feb 57 158 51 48 59 53Feb 57 – Jun 70 160 64 50 46 53

Jun 70 – Sep 81 135 51 41 43 45

Sep 81 – Nov 90 110 32 44 34 37

Nov 90 – Oct 01 131 44 42 45 44

Averages 139 48 45 45 46

Table 14.2 The four-year cycle in the Dow

low might be seen as being caused by a rightward-biased sub-cyclethat inevitably had to succumb to the magnetic pull of a cycle low.

THE 1987 CRASH

A very clear example of this phenomenon occurred in the Australianequity market during the 1987 Equity Crash. Figure 14.13 shows theone-year percentage rate of change in the Australian All OrdinariesIndex between late 1974 and late 1989. Looking just at the registeredlows in the index, it would have been very easy to conclude that therehad been two beats of a seven-year cycle, which had therebycompleted a 14-year cycle. The conclusion would have been givencredence by the fact that not only did the cycle begin and end after


Figure 14.12 Momentum and the energy gap

significant bear phases (ie, the 1973–74 bear and the aftershocks fromthe 1987 Crash) but both beats of the apparent seven-year cycle ‘look’the same.

However, let us suppose that the performance of the All OrdinariesIndex can be redefined to allow for the possibility that the 1987 Crashwas some form of energy gap. And let us use a three-monthpercentage rate of change indicator to enable us to focus more directlyon short-term cycles. Figure 14.14 shows the resulting momentum inthe All Ordinaries Index immediately prior to, and just after, theCrash. It can be seen that the high of 25 September 1987 occurs at themid-point of what could be a 6.33-year cycle. Moreover, this 6.33-year cycle clearly sub-divides into three 2.11-year cycles.

A complete analysis would, of course, need to extend forward andbackwards in time, in order to ensure that both the alleged 6.33-yearcycle and the underlying 2.11-year cycles repeated themselves.Nevertheless, the example quite clearly indicates that using an energygap to isolate a potential cycle peak, and then searching forward andbackwards in time for momentum lows that are balanced around thatpeak, can produce clear evidence of six-wave cyclical fluctuations.This is a more satisfactory outcome than assuming that the low after asharp market fall marks the end of a cycle. In fact, a confirmed energygap is prima-facie evidence, not of a cycle low, but of a cycle peak. Itis then a simple matter to find lows prior to the peak and lows after the


-60

-40

-20

0

20

40

60

80

100

120

Se

p-7

4

Ma

r-7

5

Se

p-7

5

Ma

r-7

6

Se

p-7

6

Ma

r-7

7

Se

p-7

7

Ma

r-7

8

Se

p-7

8

Ma

r-7

9

Se

p-7

9

Ma

r-8

0

Se

p-8

0

Ma

r-8

1

Se

p-8

1

Ma

r-8

2

Se

p-8

2

Ma

r-8

3

Se

p-8

3

Ma

r-8

4

Se

p-8

4

Ma

r-8

5

Se

p-8

5

Ma

r-8

6

Se

p-8

6

Ma

r-8

7

Se

p-8

7

Ma

r-8

8

Se

p-8

8

Ma

r-8

9

Se

p-8

9

1-ye

ar %

ch

ang

e

Australian All Ordinaries Index

1-year % change on daily closes(Data source: Datastream)

Equity Crash

Figure 14.13 Dyadic cycles and the 1987 Equity Crash

peak that are equidistant from the peak. The time elapse between anysuch pairs of lows will then provide an indication of what cycle peri-odicities are operating.

CONFIRMATION OF THE ENERGY GAP

When an energy gap actually occurs in a market, people know itbecause they experience it. However, when researching historicaldata, to establish the rhythm of a cycle, it is not possible to accessemotions directly. We need a proxy of some sort. Quite obviously, anacceleration in momentum (whether up or down) is a good startingpoint. However, we also need an indicator that can provide inde-pendent confirmation. One such indicator is the percentage intra-dayspread between price highs and lows. The important point here is thatan energy gap implies an underlying imbalance between supply anddemand. In extreme cases, once a market has entered a ‘bubble’period, and demand literally becomes one-way, a point will arrivewhen long-term fundamental investors are satiated and only a smallamount of short-term technical selling will burst the bubble. In otherwords, selling will have an unusually strong downward impact onprices, because there are no natural buyers into weakness. Thisimportant moment can usually be identified by a significant move inthe spread between the daily high and the daily low in prices.


-50

-40

-30

-20

-10

0

10

20

30

40

50

Ju

l-8

4

Oct-

84

Ja

n-8

5

Ap

r-8

5

Ju

l-8

5

Oct-

85

Ja

n-8

6

Ap

r-8

6

Ju

l-8

6

Oct-

86

Ja

n-8

7

Ap

r-8

7

Ju

l-8

7

Oct-

87

Ja

n-8

8

Ap

r-8

8

Ju

l-8

8

Oct-

88

Ja

n-8

9

Ap

r-8

9

Ju

l-8

9

Oct-

89

Ja

n-9

0

Ap

r-9

0

Ju

l-9

0

Oct-

90

3-m

on

th %

ch

ang

eAustralian All Ordinaries Index

3-month % change on daily closes(Data source: Datastream)

Equity Crash

Figure 14.14 Triadic cycles and the 1987 Equity Crash

THE WALL STREET CRASH

One of the best examples is the Wall Street Crash of 1929. The start ofthe Crash, as reflected in the Dow Jones Industrial Average, is shownin Figure 14.15. Here daily closes on the Dow are correlated with theintra-day percentage spread between the index high and the indexlow. If the day is an up day the spread is positive; if it is a down day,the spread is negative. The Dow reached a closing peak for the periodon 3 September 1929. It then began to fall away. There were shortperiods, when the index fell for a number of consecutive days. Thiswas certainly unusual in the context of what had gone before, and wastherefore a warning sign; but the daily price spread remained rela-tively unaffected. Note, however, the way that the daily spreadballooned out on 23 and 24 October 1929, and then did so again on 28and 29 October. This was the energy gap.

As predicted by theory, at some stage after the market had nego-tiated the energy gap (and implicitly completed its task of reversingthe polarity of the market), prices began to rally. In fact, after theOctober 1929 shock, the rally on the Dow lasted until mid-April1930, thereby lulling investors into a false sense of security. And thenthe impulsive down-wave materialized. In fact, the intra-day spreaddid widen out as the market tipped into that down-wave but it nevercame anywhere near the levels experienced during the energy gap.


175

200

225

250

275

300

325

350

375

400

01-J

un-2

9

11-J

un-2

9

21-J

un-2

9

01-J

ul-2

9

11-J

ul-2

9

21-J

ul-2

9

31-J

ul-2

9

10-A

ug-2

9

20-A

ug-2

9

30-A

ug-2

9

09-S

ep-2

9

19-S

ep-2

9

29-S

ep-2

9

09-O

ct-2

9

19-O

ct-2

9

29-O

ct-2

9

08-N

ov-2

9

18-N

ov-2

9

28-N

ov-2

9

08-D

ec-2

9

18-D

ec-2

9

28-D

ec-2

9

07-J

an-3

0

17-J

an-3

0

27-J

an-3

0

06-F

eb-3

0

I

ndex

leve

l

-16

-12

-8

-4

0

4

8

12

16

Hi/L

o %

spr

ead

Dow Jones Industrial Average(Daily, high/low/close)


DJ Index(l.h.scale)

Hi/Lo % spread(r.h.scale)

Figure 14.15 Confirmation of the energy gap

CYCLE CHARACTERISTICS

The fact that cycles cluster in triads suggests that each cycle mayhave a different purpose, which is in some way related to theevolution of the higher-level cycle of which it is a part. If this iscorrect, then it is also likely that each cycle will have specific ‘foot-prints’ that may help to identify it. There will always be significantvariations, based on the essential creativity of markets. Nevertheless,certain practical generalizations are possible, particularly in relationto cycles that last for months and years. These generalizations relateto cycle functions, cycle biases and cycle behavioural traits. In theanalysis that follows, it is assumed that the higher-level cycle is onethat lasts for about 11½ years and that the lower-level sub-cycles lastfor about 46 months. This is consistent with the data in Tables 14.1and 14.2.

CYCLE FUNCTIONS

The first set of generalizations relate directly to the position of a cyclein a triad (see Figure 14.16). This is, in a sense, an extension of thediagram that dealt with the momentum distortions created by energygaps (Figure 14.12). The first cycle in a triad is likely to encompassthe information shock, or energy gap, that initiates the higher-levelcycle. It is also likely to encompass the initial re-test of the low. Inother words, the first cycle embraces the period when the market


Figure 14.16 Cycle functions within a triad

learns that the trend has changed. As such, it is the base cycle withinthe context of the higher-level cycle.

The second cycle in a triad is likely to be quite dynamic during itsrally phase. The market will have learnt that a reversal has occurred andwill be in a natural impulse wave. At the top of the cycle, of course, themarket will be genuinely overbought. Consequently, the down phase ofthe second cycle will likely encompass the information shock, or energygap, that initiates the bear phase. The second cycle is therefore the trendcycle, although it also incorporates the start of a reversal.

The third cycle in a triad fulfils the task that follows on from this. Itsrally phase will incorporate the re-test of the high as the market learnsthat a reversal has occurred. However, its down phase will involve themain thrust of the bear. The third cycle in a triad, therefore, is theterminal cycle, which unwinds the excesses of the previous cycle andleaves the market oversold.

CYCLE TRANSLATION

The second set of generalizations relating to cycles arises directlyfrom each cycle’s implicit relationship to the higher-level cycle ofwhich it is a part. Most analysis will be done on the assumption that acycle will peak somewhere near the centre of its periodicity. Hence,for example, a 44-month cycle might be expected to peak somewherebetween months 21 and 23. However, many analysts have noted that –in contrast to cycle lows, which tend to arrive (more or less) in anexpected time frame – cycle highs can arrive significantly earlier orlater than expected. In other words, there is more variability in thetiming of peaks. This is why it is always best to measure cycle period-icities in relation to low points.

Deviations from the central timings of cycle highs are called ‘trans-lations’. A left translation means that a cycle peaks significantlyearlier than the middle date, while a right translation means that acycle peaks significantly later.

TRANSLATIONS WITHIN A TRIAD

The effects of a higher-level cycle on a lower-level triad of cycles areshown in Figure 14.17. The first cycle, which initiates the bull phase,is likely to be the subject of uncertainty and to be characterized by


slow learning. It is unlikely that it will be characterized by any greatdegree of urgency in either direction, up or down. In general,therefore, a base cycle is likely to be neutral in relation to its centrepoint, or unbiased.

The second cycle, which is characterized by bullish sentiment, islikely to incorporate the long haul into an overbought condition. Atsome stage, new money is no longer available from cash reserves; itwill either have to be borrowed or earned before it can be committedto the market. Also, the influx of bullish data, which confirm the trend,will be spread over time. These two forces suggest that a trend cycle islikely to peak late: it is biased to the right, or is right-translated.

The third cycle in a triad is destined to unwind the excesses of thetrend cycle. The fall at the end of the trend cycle will already havewarned that the market was overbought and investor sentiment willhave started to shift from the fear of being left out of the market to thefear of being left in the market. This shift will ensure that each rallywill tend to be relatively limited in price and that falls are extended intime. Consequently, a terminal cycle is likely to peak early. Ittherefore tends to be left-translated.

THE INFLUENCE OF HIGHER-LEVEL CYCLES

As shown in Figure 14.17, the higher-level cycle (represented by thedotted line) is implicitly assumed to be passive in relation to the biases


Figure 14.17 Cycle translations within a triad

in its lower-level constituent cycles. However, this is far from beingthe case. On the basis of what has been noted about the different func-tions of different cycles, it is possible to hypothesize that translation ofa specific level depends on two factors – namely, the position of thecycle within its triad and the influence of the higher-level cycle.

The distinction between the two influences is important. It cannotbe emphasized enough that the bias of a lower-level cycle is deter-mined by the influence of its immediate higher-level cycle; and thatthe latter depends on fundamental trends in the socio-economic envi-ronment. This is one of the reasons why it is so dangerous to ignore‘fundamentals’. We shall accordingly be looking at economic cyclesin Chapter 15.

Some of these issues are shown in Figure 14.18. Here, the sequenceof biases shown in Figure 14.17 is reflected in the likely biases oflower-level cycles. Hence, the triad in the base cycle is likely to becentred, and the triad in the trend cycle is likely to reproduce the biasesof all three higher-level cycles.The triad in the terminal cycle is oftenless certain. The middle sub-cycle, for example, might reflect a long,slow struggle against the downtrend (ie, be right-translated) or it mightbe characterized by a short rally and a long fall (ie, be left-translated).Nevertheless, the first of the three should be centred, the second shouldallow a recovery that is limited in price but long in time, and the thirdshould reflect the final capitulation.

CYCLE BIASES AND ENERGY GAPS

There is an underlying issue here about where precisely the criticalenergy gap – which reverses the polarity of a cycle from bullish to


Figure 14.18 Lower-level cycle translations

bearish – will actually impact. If the trend cycle is particularlystrong, and particularly rightward-biased, the late bear phase maynot actually reverse the market’s polarity (although it will warn ofsuch a reversal). This will allow the market to make a new high inthe first sub-cycle of the terminal cycle and, as a result, the energygap will impact when that first cycle turns down. Often the onlyway to tell whether or not an energy gap has occurred is to look atthe energy indicators.

CYCLE BEHAVIOURAL TRAITS

The third set of generalizations relating to cycle characteristics is thateach cycle within a triad will reflect certain behavioural traits byparticipants. It is as if the attitudes of the many become so homoge-nized and condensed that they can be analysed as if they were beingexhibited by just one person. This is one of the implications of group,or crowd, behaviour. As we indicated in Chapter 2, an effective crowdgains critical mass through conformity enforcement, therebysuppressing diversity generation. Hence, the capabilities of indi-viduals within the crowd are reduced down to a common level.

In theory, all behaviour has three dimensions to it – namely,affective, somatic and cognitive.4 With a crowd the quality of thesedimensions will essentially be very simplistic, not least because theywill be orientated towards fulfilling the crowd’s purpose. We have justseen that each cycle within a triad has a specific purpose. So the crowdinvolved with the cycle will likely exhibit a blend of characteristicsthat reflect that purpose. Indeed, that blend will be the one that helpsto bind the participating individuals together. For each of the cycles ina long-term triad, we can hypothesize that the emotional tone, thethrust of the activity and the tendency in thought processes will be asshown in Figure 14.19.

Sometimes, these traits will be easy to identify by observation andintrospection. If this is so, analysts can have an even clearer under-standing of what is happening. Often, however, they may not be easyto identify, particularly over short time horizons. In this case, it mayonly be possible to guess at, or intuit, them. Nevertheless, a generalunderstanding of the psychosomatic basis of the crowd can providea distinctive aid to understanding the likely consequences forfinance, business and politics. So let us look briefly at each of thecycles in turn.


BEHAVIOURAL TRAITS IN A BASE CYCLE

Each cycle in a triad is very much influenced by – and is, indeed, aderivative of – the cycle that preceded it. Hence the first cycle in a triadis characterized by a reaction to the preceding crisis from which it wasborn. It is the most passive cycle because there is a need to recoverfrom the shock. Feelings will be slow-moving and indolent; activity interms of market transactions will be sporadic and short-lived and willreflect the overall need for rest and recuperation; and, importantly, thecognitive bias, or the mental lens through which the world is seen, willbe clouded by a denial that circumstances are moving beyond a crisisatmosphere. This combination is essentially very dangerous because,under certain circumstances, despair may set in.

BEHAVIOURAL TRAITS IN A TREND CYCLE

The second cycle is the most creative. It is the one that incorporates thelearning generated from the base cycle. It exhibits a recognition thatcircumstances have actually changed. Feelings are therefore likely tobe coloured by a sense of confidence about the trend, both now and inthe future; activity is likely to be orientated towards change and will bethe more dynamic for it; but, strangely, the cognitive bias will incor-porate a sense of deception inasmuch as attention increasingly focuseson market price movement rather than underlying fundamentals. Inother words, during a trend cycle, sentiment and activity can becomeincreasingly divorced from reality.


Figure 14.19 Behavioural traits within a triad

BEHAVIOURAL TRAITS IN ATERMINAL CYCLE

The third cycle within the triad is overtly destructive. Whereas setbacksin the previous two cycles may have been quickly reversed, in theterminal cycle they may translate into something more dramatic. Inother words, the setback has the quality of a fundamental bear ratherthan of a technical correction. So it is no surprise that the dominantfeeling tone of the third cycle is one of fear.5 In the face of a genuinethreat, the only valid responses are to fight, to flee or to freeze. Someinvestors will basically fight the bear by buying into weakness. Theirreaction is primarily emotion-based and will, more often than not,involve an element of anger.6 Some investors will run from the bear byselling. Their reaction can best be described as body-based since theprimary motivation is physical protection.7 The vast majority, however,will be unable either to fight or to flee. They will have been shocked bythe energy gap (which may have occurred at the tail end of the trendcycle or in the first downturn of the terminal cycle) and they willfreeze.8 The somatic reaction of the crowd will therefore be one of adeep-rooted resistance to the actual ending of the trend cycle.9

Interestingly, under these circumstances, there will be a naturaltendency to blame others, or external events, for the undesired changes.The cognitive bias of the third cycle in a triad, therefore, will be one ofaccusation. There is, however, one consolation from this cycle. This isthat, after initial shock, fear is likely to eventually translate into action.This action is explicitly focused on avoiding further damage but is alsoimplicitly orientated towards trying to return to the wellbeing of theprevious cycle. As a result, the third cycle has a tendency to be shorterthan the previous pair.

A SCHEMATIC FOR FINANCIAL MARKETS

We can pull these various strands together and express them in theform of a schematic diagram (see Figure 14.20). This diagram is basi-cally an adjustment of that shown in Figure 14.6. It takes into accountthe potential biases – devolved from the relevant higher-level cycle –that might be experienced by the lower-level cycles. Thus, the first(base) cycle is centred; the second (trend) cycle is biased to the right;and the third (terminal) cycle is biased to the left. The result is a patternthat hints at three types of corrective move: a deep one at the end of the


first cycle; a short, sharp, one at the end of the second cycle; and twomore within the third cycle that are interspersed with a rally. Most (ie,somewhere between two-thirds and five-sixths) of the third cycle thuspresents itself as a correction to the whole pattern.

Interestingly, this schematic is a close reproduction of the patternshown in Figure 14.5, which showed the six-month percentage changein the Dow Jones Industrial Average between November 1990 andOctober 2001. As will be demonstrated in Chapter 17, this period wasdominated by revolutionary innovation. The pattern shown in Figure14.20 is therefore one that implies change within the system. Thepattern will reappear again in our analysis in due course.

LIMITED CYCLE PATTERNSThis analysis indicates that any particular cycle can only present itselfas one of a limited number of patterns, based around an underlyingthree-up/three-down process: a balanced form, an upward rising formor a corrective form. This further means that properly specified triadsof cycles will also present in the same limited number of ways.Indeed, for the bigger cycles the pattern is likely to look somethinglike that shown in Figure 14.20.

CONCLUSION

We saw in the previous chapter that the intrinsic pattern within acyclical upswing or downswing is the three-wave pattern of recog-


Figure 14.20 The cycle pattern in financial markets

nizing, absorbing and responding to changes in the socio-economicenvironment. In this chapter we have extended the analysis to showthat cycles cluster in groups of three – that is, there are three sub-cycles in a higher-level cycle, each of which has a different purpose inrelation to that higher-level cycle. Further, each sub-cycle is likely toreach a peak that bears a very specific relationship to the theoreticalcentre point of the cycle. Finally, the resulting bias almost certainlyreflects a specific combination of subjective qualities based on thethree characteristics of mood, ‘physicality’ and cognition.

There is thus an essential ‘threesomeness’ to cycles that manifestsitself in a number of dimensions. This is further evidence of the non-random nature of the phenomenon of markets.

NOTES

1. There is also the added point that the Dow Jones Industrial Average is partneredby the Dow Jones Transportation Average. The two indices are often trackedtogether to establish the degree to which they are behaving in a similar fashion.This is the basis of ‘confirmation’ and ‘non-confirmation’.

2. See Wilson, Louise (1964) Catalogue of Cycles: Part I – Economics,Foundation for the Study of Cycles, Pittsburgh.

3. Importantly, confirmation of a (roughly) 46-month cycle comes from othersources. First, it corresponds to the length of the short-term business cyclefound by Joseph Kitchin. See Kitchin, Joseph (1923) ‘Cycles and trends ineconomic factors’, Review of Economic Statistics, 5, Harvard EconomicService, Cambridge, Massachusetts. Second, and more directly, it falls into therange of periodicities between 40 months and 50 months that have beenreported by Leonard Ayres, Ellsworth Huntingdon and Edward Dewey. SeeAyres, Leonard (1939) Turning Points in Business Cycles, Macmillan, NewYork; Huntingdon, Ellsworth (1941) ‘Effect of atmosphere electricity onbusiness’, in The Frontier, Armour Research Foundation, Chicago; andDewey, Edward (1955) ‘The 3½-year cycle in general business’, in Cycles, VI,Foundation for the Study of Cycles, Pittsburgh. For a complete list, see Wilson,Louise (1964) Catalogue of Cycles: Part I – Economics, Foundation for theStudy of Cycles, Pittsburgh.

4. The characteristics are derived from the work of psychologist Claudio Naranjo.Naranjo argues that there are three basic centres – affective, physical andcognitive – in each person, around which that person’s character structureconstellates. Since these centres operate in combination, and since each combi-nation has a different hierarchical pattern, the result is that there are ninepotential centres of gravity within the psyche. Each person will operate fromone of those gravity centres. See Naranjo, C (1994) Character and Neurosis: Anintegrative view, Gateways/IDHHB, Nevada City, California. Some of theimplications of Naranjo’s work for individuals are explored in Chapter 25. The implications for a crowd are much simpler. All that we need to know are thebasic affective, somatic and cognitive characteristics of birth, growth and decay.


5. The higher the level at which the cycle is operating, the more pronounced andpervasive the feeling of fear will be.

6. The anger is likely to arise because the market has confronted the investor’sbeliefs about himself or herself. See also Chapter 25.

7. The physical bias in ‘running’ from the market is very important. Marketprofessionals sometimes refer to it as ‘gut feel’.

8. The ‘freeze’ response is the one that is least understood by most people,including psychologists, and yet is the one that is most widely experienced.However, it is the most likely outcome when a person can neither fight nor flee.The resulting shock, or trauma, becomes locked in the psychosomatic systemand can easily be activated (and re-lived) by apparently non-relevant events.Post-traumatic stress disorder is only the most extreme example of this. Someanalysts argue that traumatic shock is a universal malady. See, for example,Levine, Peter A (1997) Waking the Tiger: Healing trauma, North AtlanticBooks, Berkeley, California.

9. This, of course, will ultimately be the primary reason why the bear will take along time to resolve itself and why downward price movements will be vicious.


201

15

Economic cycles

INTRODUCTION

Technical analysis seeks to be objective, basing decisions on theevidence that is directly available from the markets themselves. Littleor no account is therefore taken of the subjective interpretation ofeconomic and social trends. Indeed, there is strong evidence that theattempt by investors to anticipate the future means that market pricesstart to turn before actual fundamentals. There is, for example,convincing evidence that major lows in the US equity marketsprecede major lows in the US economy by between four and sixmonths.1 Under these circumstances, there seems little point inanalysing fundamentals.

Nevertheless, there is an aspect of the relationship between marketsand the economy that needs to be taken into account. This is the factthat there comes a point in a bull or bear trend when the markets andfundamentals enter a strong feedback relationship with each other. Itis the presence of feedback that generates the trend in the first place.For example, there comes a point where falling equity prices impactwealth and confidence to the extent that consumption and investmentstart to contract. Consequently, economic activity slows, therebyincreasing unemployment, reducing spending power and furtherundermining confidence. So equity prices fall. And so on. Technicalanalysis is not necessarily very good at determining when thisfeedback process actually begins. This raises the question about whattechniques might be used to address this deficiency.

There are two methods that can be used. The first, which will beaddressed in Chapter 22, is the application of the golden ratio(0.382:0.618) to identify critical boundaries. The second, which is thesubject of this chapter, is to apply the techniques of cyclical analysis tothe economy itself. If we can objectively know the current status ofthe economy within the big scheme of things, then we have some indi-cation of the implicit bias in a financial market cycle, we have animproved chance of anticipating official policy adjustments thatmight impinge on a market and we have a good chance of beinginvested the right way during a trend move.

NOTE ON ECONOMIC THEORY

Mainstream economic theory basically discounts the inevitability ofeconomic cycles. There are three reasons for this. The first, which hasalready been alluded to in Chapter 6, is that economic theoryexcludes, by assumption, the controlling influence of the crowd. Ineconomics, the rational individual reigns supreme and – like a ClintEastwood character in a spaghetti western – does not take muchaccount of the behaviour of others. Further, such rational behaviour islinear. Under these circumstances, economic fluctuations are essen-tially caused by unforeseen (exogenous) ‘shocks’.

The second reason why economic theory does not see economiccycles as being inevitable follows from this. If economic fluctuationsare in some way an accident, then a government can use its authorityand resources to offset any undesirable developments. Hence agovernment is empowered to lower interest rates, reduce taxes andincrease government spending in a recession and it is empowered toraise interest rates, increase taxes and reduce government spendingunder conditions of inflation. In theory, it is possible to offseteconomic fluctuations and converge on non-inflationary/highemployment trend growth.

The third reason why economic theory has problems in recognizingthe influence of economic cycles is that the historical evidence doesnot support constant periodicities for cycles. In 1923, Joseph Kitchinreported a short-term, three- to five-year, business cycle based oninventory adjustments.2 In 1862, Clement Juglar reported a medium-term, seven- to 11-year, cycle relating to capital investment.3 In 1930,Simon Kuznets pointed to the idea of a 15- to 25-year cycle involvingthe interaction of construction and demographic factors.4 And thereare even longer cycles on offer. In 1991, Brian Berry pointed to the


presence of a generation-length, 25- to 35-year, cycle that was relatedto investment in infrastructure.5 In 1926, Nikolai Kondratyevpublished his classic article on long waves of 40 to 60 years incommodity prices.6 Finally, in 1991, social economists WilliamStrauss and Neil Howe presented a strong case for an inter-generational, 85- to 99-year, crisis cycle.7 However, despite the depthof research conducted by these economists, there is a tendency toignore the role of cycles. There are two basic reasons for this. First, itis very difficult to identify appropriate causal agents, particularlywithin long-term cycles. Second, the variation in periodicities makesit difficult to use cycles for forecasting purposes. It is therefore mucheasier to ignore the research and concentrate instead on short-termcontrol mechanisms.

AN INTEGRATIVE VIEW

The fact remains, however, that certain important cycles have beenshown to exist. Suppose, therefore, that we could find the presence ofa three-wave adjustment mechanism within these cycles. This mech-anism – of information shock, information absorption and infor-mation application – would be endogenous to the process and wouldnot necessarily rely on an external timing mechanism. Nor would itnecessarily require a specific internal causal mechanism. Wouldn’tthis mean that the pattern of the cycle would be at least as importantas the periodicity of a cycle?

If this is correct, then the correct approach to cycles is to search, notjust for specific cycle periodicities, but instead search for completingpatterns within time windows that are defined by the variability of theperiodicities. Indeed, we could even compare currently evolvingeconomic cycles with previous completed cycles. If we thus knowwhere we are on the economic cycle, we can have some idea aboutwhat to expect from markets. This would have the potential to signif-icantly reduce forecasting uncertainties.

Importantly, this integrative approach allows both for creativity ineconomic processes and for some form of constraining order in thoseprocesses. On the one hand, the intrinsic creativity of human actiongenerates variability in the periodicity of a cycle. On the other hand,the natural processes of human life ensure that the evolution of thatcycle will conform to an underlying pattern. Hence, randomness andorder co-exist.

Economic cycles 203

RELATIONSHIPS BETWEEN ECONOMICCYCLES

Let us start the analysis by looking briefly at the possible timing rela-tionships between the economic cycles that are indicated above. Thefirst important step here is to incorporate the research findings of otherindependent analysts to see if there are any obvious ‘averages’ thatcan be used. The second important step (given the analysis of theprevious chapter) is then to look specifically for cycles that aregrouped in batches of three. Obviously, this is something of an iter-ative process. The basic idea is to arrive at a set of cycle averages thatincorporates all the main cycles in triads.

FROM KITCHIN TO STRAUSS AND HOWE

There is a huge amount of evidence that the central periodicity of theshort-term Kitchin cycle (or inventory adjustment cycle) is some-where between 40 and 44 months – that is, somewhere between 3.33and 3.67 years.8 These periodicities can be found in indices of generalbusiness activity and in indices relating to specific industrial sectors.There are variations depending on the length of data that is analysedand on the purposes of the researcher; and there are always outliersthat can distort the evidence. Nevertheless, this is the one cycle lengththat, if pressed, most economists would accept as being correct.

If we then multiply this short-term business cycle by three, we findthat the next higher-level cycle has a periodicity of 10 to 11 years. Thisis undoubtedly the potential range for the Juglar cycle (or capitalinvestment cycle). Multiplying this by three gives an outcome of 30 to33 years. This is the range for the Berry cycle (or infrastructure cycle).Multiplying the Berry cycle by three results in 90 to 99 years, whichcoincides with the Strauss and Howe cycle (or crisis cycle).

FROM STRAUSS AND HOWE TO KITCHIN

Now, Strauss and Howe’s findings are quite specific about the natureof the 90-year crisis cycle. On the basis of an analysis that goes backto the 16th century, they argue that ‘social moments’ occur every 45years.9 A social moment is defined as being either a secular


(economic) crisis or a spiritual (social) awakening. Hence, there is asecular crisis every 90 years. If we use this as our starting point, andnow divide by three, we get 30 years, which is the Berry infrastructurecycle. If we divide this by three, we get 10 years, which is the Juglarinvestment cycle. And if we divide this by three, we arrive at 3.33years, which is the Kitchin inventory cycle.

The basic outline of these simple relationships is shown in Table15.1. There is no question that we are looking at a series of cycles withcentral periodicities of 3.33 years, 10 years, 30 years and 90 years.

KUZNETS AND KONDRATYEV

There are, however, two cycles that need further comment. These arethe 20- to 25-year Kuznets cycle and the 45- to 60-year Kondratyevcycle.

The Kuznets cycle is no longer generally accepted as being aseparate cycle.10 A 20-year cycle was very evident prior to the FirstWorld War, but has not obviously been present since. However, BrianBerry has counter-argued that a Kuznets-type cycle does exist, but thatit is longer than Kuznets thought. That is, its central periodicity is 30years, rather than 20 years. Berry calls his 30-year cycle a Kuznetscycle. Actually, to avoid confusion and to take account of the fact thatBerry shifts the emphasis from building to general infrastructure(which includes intellectual infrastructure), it is more accurate to callthe 30-year cycle the ‘Berry’ cycle.

Economic cycles 205

Table 15.1 Relationships between economic cycles

Researcher Period (years) Average Relationships

Kitchin 3–5 3.33

Juglar 7–11 10 3 × Kitchin

Kuznets 20–25

Berry 25–35 30 3 × Juglar/9 × Kitchin

Kondratyev 45–60

Strauss and Howe 85–99 90 3 × Berry/9 × Juglar

The Kondratyev cycle is a price cycle, not an economic cycle as such.This is an important distinction, which is often either ignored orforgotten. Kondratyev saw prices as rising and falling in long waves. Itis therefore appropriate to see which groupings of economic cyclescoincide with one complete beat of a Kondratyev cycle. We shall look atthis in Chapter 17. In the meantime, we can point to some interestingrelationships between the Kondratyev cycle and the Strauss and Howecrisis cycle. The first conclusion is that not all Kondratyev lows aremajor depressions because not all Strauss and Howe lows will coincidewith a Kondratyev low. The second, which follows from this, is thatStrauss and Howe crises will alternate between deflation (persistentlyfalling prices) and accelerating inflation (rapidly rising prices).

ECONOMIC THEORY AND TECHNICALANALYSIS

The above survey hardly does justice to the massive amount ofresearch and analysis conducted by the specifically named individualsand nor does it include the wealth of supportive information providedby other analysts. Further, the analysis has been concentrated on aparticularly narrow range of cycle periodicities. However, the purposeof this exercise is not to embark on a critique of economic causationbut to point out that a) impeccable research by respected economistsconfirms that certain important economic cycles definitely exist andthat b) a simple examination of the periodicities indicates that thesecycles are probably related. In other words, there is evidence thateconomic cycles are ordered processes and that these processes seemto adhere to an intrinsic ‘rule of threesomeness’.

In a sense, this is all that a technical analyst needs to know. Once thedecision has been taken to pursue an as-objective-as-possibleapproach to market timing, an analyst need only track the evolution ofa cycle’s momentum through time. The underlying economic theoryand political nuances become irrelevant. It is not necessary, forexample, to understand in detail the precise theoretical mechanismsbehind a 60-year, or 90-year, cycle. It is only necessary to know thatthere is strong evidence that they do, in fact, exist.

In addition, since the analysts will be concentrating on behaviouralpatterns it is not necessary to use sophisticated statistical techniques tocalculate either the precise periodicity, or the precise historical vari-ability, of cycles. All we need to know is that a cycle series with


approximate, but inter-related, periodicities is operating. Hence, in thecycle series outlined above, reference to a ‘3.33-year cycle’ or a ‘10-year cycle’ does not mean that those periodicities are precise and invi-olable. The references are merely labels that point us in the directionof an important phenomenon.

CYCLE CHARACTERISTICS

Before looking at some historical data in any detail, there is one morepiece of the theoretical jigsaw that needs to be put in place. This is thefact that each economic cycle within a triad is likely to fulfil adifferent purpose to its neighbours. This is exactly the same point aswas made in Chapter 14 about financial market cycles.

Let us look briefly at the likely characteristics of a longer-termcycle (without being too precise about the definition of ‘longer-term’).There are a number of points that can be made:

� each new cycle is born from the downswing of the previous one;

� the overarching longer-term cycle will begin from a recession andwill end with a recession;

� the first sub-cycle (the base cycle) in a triad will therefore beinvolved in re-establishing some form of stability in the economy;

� the second sub-cycle (the trend cycle) in a triad will incorporateany qualitative changes that are appropriate after the re-basing;

� the third sub-cycle (the terminal cycle) will embrace the recessionthat concludes the higher-level cycle.

However, each sub-cycle obviously consists of an upswing and adownswing. So there will be two sub-cycle recessions between eachof the two larger-scale recessions. This is shown in Figure 15.1.

BIASES IN CYCLES

It is thus possible to hypothesize that the higher-level cycle will bequantitatively affected by the lower-level cycles, and that the lower-level sub-cycles will be qualitatively affected by the higher-levelcycle. What this basically means is that the shorter-term cycles will bethe vehicle by which longer-term structural changes in the economy

Economic cycles 207

are laid down. It also means that the fluctuations in each shorter-termcycle will be biased both in amplitude and time by evolutionary devel-opments in the longer-term cycle. Consequently, each cycle in a triadis likely to have a specific type of bias that will be unique to itself. Sowe shall now look at each cycle in turn, building on the strands ofthought that we started to develop in the last chapter.

For ease of presentation, we can present the observations graphi-cally, as in Figure 15.2. The diagram shows a notional index ofbusiness confidence as it might develop over three contiguous cycles.The dashed line shows the overarching long-term cycle. The threesub-cycles are referred to as the ‘base’ cycle, the ‘trend’ cycle and the‘terminal’ cycle respectively. These three cycles are treated as havinga similar time elapse to one another. Further, each sub-cycle is shownto consist of three lower-degree cycles.

The diagram is highly schematic inasmuch as the amplitude ofthe economic fluctuations will be more variable than indicated andthe bias in the cycles may diverge from those indicated.Nevertheless, the diagram shows clearly how a triad of cycles mightrelate to one another. The pattern is the same as the one shown inFigure 14.15.


Figure 15.1 Cycle and sub-cycle recessions

AN EXAMPLE FROM HISTORY

It might be useful at this stage to confirm that there is some practicallogic behind the diagram in Figure 15.2. Figure 15.3 accordinglyshows the Berry-length cycle in US industrial production that beganin the winter of 1946 and ended in the summer of 1980. The evolutionof the cycle is tracked in terms of two-year percentage rates of change.It is often useful to use two-year rates of change when searching forlonger-term cycles, because it helps to cut out short-term fluctuationswhile still giving a good indication of the timing of important lows.The period covers just over 34 calendar years. It was characterized byrestoration after the Second World War, by the Cold War and by theVietnam War. It was therefore a period of increasing governmentdeficit financing and rising inflation.

JUGLAR CYCLES DURING THE 1946 TO 1980BERRY CYCLE

The first Juglar cycle in this period, measured from low to low, lastedfrom February 1946 to April 1958. It was basically associated with thegradual restoration of normality after the trauma of the GreatDepression and the Second World War and was clearly characterizedby a three-wave expansion (1–2–3) and a three-wave contraction(A–B–C). The peak of wave 3 was more or less in the middle of thecycle.

Economic cycles 209

Figure 15.2 A triad of economic cycles

The second Juglar cycle in the period lasted from April 1958 toNovember 1970. It was the period of social upheaval and change thatis now called the ‘Swinging Sixties’. It was also very inflationary. Theexpansion from May 1961 to October 1966 was one of the longest onrecord. The cycle was therefore rightward biased, with an elongatedupwave (1–2–3) being followed by a short, sharp downwave(A–B–C).

The third Juglar cycle lasted from November 1970 to August1980. The first phase was an inflationary recovery, which was haltedin its tracks by a sharp rise in oil prices. The two-year percentagechange in industrial production fell sharply between August 1973and May 1975. There was then another sharp recovery into June1977 before a final collapse into June 1980. The overall profiletaken by this crisis cycle was of an initial sharp recovery followedby a slow three-wave fall.

THEORY AND FACT

We can therefore conclude from this important example that the infla-tionary 34-year Berry infrastructure cycle, which started in February1946 and ended in June 1980, consisted of three Juglar capitalinvestment cycles that lasted, respectively, for 12 years, 12 years and10 years. Moreover, the pattern of each cycle accords with the basic


-40

-30

-20

-10

0

10

20

30

40

50

Jan-

46

Jan-

48

Jan-

50

Jan-

52

Jan-

54

Jan-

56

Jan-

58

Jan-

60

Jan-

62

Jan-

64

Jan-

66

Jan-

68

Jan-

70

Jan-

72

Jan-

74

Jan-

76

Jan-

78

Jan-

80

2-ye

ar %

cha

nge

1

1 3

A

3

2C

B

B

A

2C

Peak

C

B

A

US Industrial Production(Monthly, 2-year % change)

Data source: US Federal Reserve Board

Figure 15.3 US industrial production, 1946 to 1980

model that has already been outlined – namely, the first cycle is a basecycle, the second (the ‘Swinging Sixties’) is a trend cycle and the thirdis a terminal cycle.

It is also worth mentioning that the terminal cycle was characterizedby a number of unpleasant shocks, ranging from an intensification ofthe war in Vietnam with the invasion of Cambodia in 1971, to a hike inoil prices after the outbreak of war in the Middle East in 1973.

THE FIRST CYCLE

We can now draw some further useful general conclusions about theintrinsic nature of each of the Juglar cycles in a Berry cycle. The firstJuglar cycle – the base cycle – will be heavily influenced by thetrauma from which it was born. The initial recovery from the lowwill be quantitative, rather than qualitative. That is, output andspending levels will recover, but will not be particularly orientatedtowards new goods or new processes. The recovery will, in fact, belargely driven by an inventory adjustment as producers try to realignoutput with better market conditions. In effect, the recovery presentsitself as an information ‘shock’ to the economy insofar as it startsunexpectedly.

However, the general economic environment will remain burdenedby the effects of the previous recession. Unemployment will still behigh and attitudes will be pessimistic. The over-riding need is for restand recuperation, so the emotional tone will be indolent and mentalattitudes will involve a denial that things are really improving. Underthese circumstances, there is little chance of a prolonged economicrecovery. Consequently, the peak will tend to be midway into the esti-mated periodicity of the sub-cycle. The subsequent downturn could bequite deep. It tends to be the stage when businesses start to take drasticaction on costs. Old plant and machinery is scrapped and unem-ployment rises again. This relapse corresponds to the informationabsorption stage of the learning process. The economy starts to detachitself from its old methodologies.

THE SECOND CYCLE

The second Juglar cycle – the trend cycle – within a Berry cycle is likethe mythical phoenix rising out of the ashes. Because surviving busi-nesses have taken some difficult decisions, they are able to respond

Economic cycles 211

more freely to genuine business opportunities. If there is a new productline, a new set of ideas about business practices, or new markets to beexploited, this is when it properly starts. The over-riding impulseduring this cycle is towards change. Accordingly, the feeling tone ispositive and confident and the mental attitude is towards higher risk.

Quite obviously, a positive economic environment has the potentialto establish a persistently high level of economic growth. If there isgenuine innovation occurring, the trend rate of growth in the economymay even appear to be improving. This is the period of transition to anew structure in the economy and it means that downturns are likely tobe mild and that the eventual peak of the cycle is likely to be delayed.The cycle is therefore biased to the right. Nevertheless, demand runspersistently ahead of supply, leading to a trade deficit and/or inflation.Eventually, consumer satiation sets in. Once this has happened, aprocess gap (ie, an energy gap) becomes inevitable.

Because of the rightward bias in the cycle, the downswing intorecession appears suddenly and is a ‘shock’. Interest rates, which mayhave been increased to stem excess demand, are quickly reducedagain. Government deficits may be expanded to help bring back thegood times. But the damage has been done.

THE THIRD CYCLE

The third Juglar cycle – the terminal cycle – within a Berry cycle is theone where the overspill from the innovation wave makes itself felt.Like all preceding cycles, it starts with a recovery of some sort.However, whereas the preceding upswing had a large qualitativedimension, this last upswing in the triad can only be quantitative. It isbased on demand stimulation, rather than industrial innovation. Assuch it is potentially inflationary. The underlying constraint is thatthere is now market satiation and therefore resistance to furtherchange.

This means that the lifespan of the recovery is limited. Quite apartfrom anything else, there is likely to be upward pressure on interestrates. Consequently, either there is a bond market collapse, or equitiescontinue to fall, or even some combination of both of these.Eventually, therefore, the recovery is aborted and a prolongedeconomic recession sets in. It needs hardly be said that, the greaterwere the excesses of the preceding boom, the greater will be therecession. The affective tone of this period becomes one of outrightfear and the mental tendency is to blame others for the problems. The


important part of this process, however, is that the ephemeral is sepa-rated from the enduring. And it is the latter that is fully incorporatedinto the deep structure of the economy.

EVOLUTION

This analysis of the stages within a cycle has brought into focus thedifferent stages of the evolutionary process within an economy. At itssimplest, each triad of cycles is involved in adjustments in the higher-order cycle of which they are directly a part. For example, threeKitchin inventory cycles will complete one Juglar capital investmentcycle; and the economy will in some way be different by the end of thethird Kitchin cycle. Another way of looking at this is to say that eachtriad of cycles involves the creation of a higher-order structure that ismore complex than the one it replaces. Hence, the information tech-nology that was introduced in the 1990s embraces a far more complexeconomic ‘system’ than did the mass consumption that dominated the1960s. However, the process of moving from a lower level to a higherlevel is not just a wild jump. Somehow the economy has simultane-ously to meet two objectives: it has to detach itself from dependenceon old structures so that it can identify itself with the newer, more inte-grated one; and it has to keep the best of the old structure while alsoadopting the new. So, for example, the new era of information tech-nology has been built on the foundations created by the era of massconsumption. Or, to put it another way, mass consumption becomes apart of the new era, rather than being the whole of an era.

The relevance of this to the current analysis is that the process ofadjustment to new economic structures consists of three stages:detachment from the old structure, transition to a new structure, andattachment to the enduring parts of the new structure.11 These stagesclearly correspond to the three phases of the triadic cycle model. Thebase cycle encourages detachment from the old paradigm (whateverthat may be). This detachment will normally be most clearly apparentin the economic slowdown that terminates the base cycle. The trendcycle involves the transition to the new paradigm (which usuallyinvolves a change in attitudes). Importantly, the end-cycle economicslowdown is relatively moderate because the adjustment is still ‘inprocess’. Finally, the terminal cycle is dominated by the separation ofthe ephemeral from the enduring (usually through the exigencies of arecession). It is only during the third, often very difficult stage that the‘new era’ paradigm becomes properly established. This is because the

Economic cycles 213

inevitable excesses of the transition phase have to be accepted andeliminated before the economy finally ‘knows’ within itself what partsof the changes are permanent. Prior to this, there is too often atendency for participants to believe that all aspects of the change arepermanent. Dotcom companies are a case in point. So, the transitionstage is not a permanent structure; and recessions and bear marketshave an evolutionary function.

LABELLING THE CYCLES

So far, we have simply referred to the cycles in a triad as being ‘base’,‘trend’ and ‘terminal’ cycles respectively. The advantage of doing thisis that the terms are basically neutral with respect to the level of thecycle being discussed. Nevertheless, there are also some advantages inbeing more specific about the essential nature of any particular cycle.For example, a base Kitchin cycle is different from a base Juglar cycle,which is different again to a base Berry cycle, which is different yetagain to a base Strauss and Howe cycle. Where it is appropriate to doso, we shall in future use the labels provided in Table 15.2.

Hence a 90-year Strauss and Howe cycle will start with a 30-yearperiod of social and political soul-searching, progress through a 30-year period of dynamic economic and social evolution, and terminatein great upheaval. A 30-year Berry cycle will be initiated by a period ofadaptation to changed circumstances, continue through a period ofsocial change, and end with a period of crisis. A 10-year Juglar cyclewill begin with a period of economic recuperation, respond vigorouslyto a period of innovation and finish with an economic disruption.Finally, the 3.33-year Kitchin cycle will start sluggishly, eventuallymove into an economic upswing and terminate in a recession.


Table 15.2 The labelling of economic cycles

Cycle Base Trend Terminal

Strauss and Howe Involution Evolution Revolution

Berry Adaptation Regeneration Crisis

Juglar Recuperation Innovation Disruption

Kitchin Recovery Advance Recession

These labels are, of course, indicative only. They pick out what arelikely to be the main features of any given period. As alreadymentioned, however, the lower levels will be strongly influenced bydevelopments at the higher levels. Consequently, some of the subtletiesof any given situation might be missed. For example, the fluctuationsof the Kitchin cycle are likely to be distorted by a conjunction ofterminal cycles at the Strauss and Howe, Berry, and Juglar levels.Alternatively, the Kitchin cycle might be very easy to pick out duringmore benign periods, such as might occur during a conjunction of basecycles at the Strauss and Howe, Berry and Juglar levels.

CONCLUSION

This chapter has established a theoretical framework for looking ateconomic cycles, based on the comments that were initially exploredin the context of financial markets in Chapter 14. We started bylooking at those cycles that respected economists have identified asbeing important. These cycles are the 3.33-year Kitchin inventorycycle, the 10-year Juglar capital investment cycle, the 30-year Berryinfrastructure cycle and the 90-year Strauss and Howe crisis, or meta-cycle. The initial evidence was that the cycles harmonize with oneanother in groups of three.

We took a preliminary look at the experience of the inflationary1946 to 1980 period in the United States. The facts seem to accordwith the theory. If this relationship is also true for other time dimen-sions, then we can draw some profound conclusions about the natureof economic oscillations. Specifically, they are ordered rather thanrandom; they are more predictable than is usually supposed; and thispredictability applies to longer-term cycles as well as to short-termones. We now need to look at some more evidence.

NOTES

1. We shall look at some of this evidence in Chapter 18.2. Kitchin, Joseph (1923) ‘Cycles and trends in economic factors’, Review of

Economic Statistics, 5, pp 10–16.3. Juglar, Clement (1862) Des crises commerciales et leur retour périodique en

France, en Angleterre et aux Etats Unis, Librairie Guillaumin et Cie, Paris.4. Kuznets, Simon (1930) Secular Movements in Production and Prices,

Houghton Mifflin, Boston.

Economic cycles 215

5. Berry, Brian J L (1991) Long-Wave Rhythms in Economic Development andPolitical Behavior, Johns Hopkins, Baltimore.

6. Kondratyev, Nikolai (1926) ‘Die langen Wellen der Konjunktur’, Archiv furSozialwissenschaft und Sozialpolitik, 56, pp 573–609.

7. Strauss, William and Howe, Neil (1991) Generations: The history of America’sfuture, 1584 to 2069, William Morrow, New York. Strauss and Howe argue thatsecular crises alternate with ‘spiritual awakenings’ every 45 years. Hence thereis a secular (economic) crisis every 90 years or so.

8. See Wilson, Louise (1964) Catalogue of Cycles, Part I – Economics,Foundation for the Study of Cycles, Pittsburgh.

9. Actually, the cycle is the length of two ‘life phases’, where a life phase is 22years long. Some analysts might argue for a different length for a life phase.However, the length of a life phase is not regarded as being absolutely precise.It is the general influence that counts.

10. See, for example, van Duijn, Jaap J (1983) The Long Wave in Economic Life,George Allen & Unwin, London.

11. This view of the evolutionary process in an economy is derived from KenWilber’s analysis of three-phase change. See, for example, Wilber, K (1998)The Eye of Spirit: An integral vision for a world gone slightly mad, Shambhala,Boston. Mr Wilber’s extraordinary grasp of the logic of change and devel-opment, particularly in the context of human consciousness, has been a majorinfluence on my thinking.


217

16

Recurrence in economicand financial activity

INTRODUCTION

In the previous chapter, we traced the theoretical evolution of aspecific pattern – or, more precisely, a triad of patterns – in economicactivity. If the theory is correct, then the patterns should repeat them-selves in a ‘fractal’-like fashion across all genuine cyclical timeperiods. This phenomenon is currently unknown to economic theory.

This, in turn, raises the possibility that we can compare theevolution of one beat of a particular cycle with the evolution of aprevious beat of the same cycle. Indeed, we can compare the patternthat has been traced out by a particular sub-cycle in a triad with thecomparable sub-cycle in an earlier triad.

This has revolutionary implications. If we know the history of acycle, then we can track any new evolving cycle directly against thathistory. There will always be the danger of unexpected shocks; but,within reason, we should be able to see where we are in the context ofthe whole.

ECONOMIC AND FINANCIAL MARKETCYCLES

The process of comparing one cycle with another can be applied bothto economic activity and to financial markets. The big advantage of

drawing economic activity directly into the analysis is that it reducesthe risk of making big mistakes. Specifically, if we know more or lesswhere we are in the evolution of the economy, we should have someidea about the likely implications for financial markets.

Moreover, and as we shall demonstrate in Chapter 22, the rela-tionship between markets and the economy is controlled by the goldenratio, 0.382:0.618. Specifically, the golden ratio can be used toestimate when a market is moving from a technical correction into afundamental trend. So, even if the equity market turns before theeconomy (which is usual but not inevitable), we can know what thatmarket might be trying to anticipate in terms of fundamentals. Further,we can know when a market trend is actually locking into feedbackwith those fundamentals. This latter is important because it marks thepoint where a trend starts to make itself felt.

SEARCHING FOR CYCLES

Figure 16.1 shows the two-year rates of change in US industrialproduction since just after the end of the American Civil War. The dataare annual, so monthly fluctuations are eliminated. The chart covers aperiod of 133 years.

At first glance, there is not much obvious order in the chart. It iseasy to pick out the period covering the ‘Roaring Twenties’, the Great


-30

-20

-10

0

10

20

30

40

50

60

70

18

68

18

73

18

78

18

83

18

88

18

93

18

98

19

03

19

08

19

13

19

18

19

23

19

28

19

33

19

38

19

43

19

48

19

53

19

58

19

63

19

68

19

73

19

78

19

83

19

88

19

93

19

98

2-ye

ar %

ch

ang

e

US Industrial Production(Annual, monthly averages, 2-year % changes)

Data sources: NBER & FRB

Figure 16.1 US industrial production, 1868 to 2001

Depression and World War II because the fluctuations are so massive.However, it is difficult to pick out a consistent periodicity betweenmomentum lows. For example, it is certainly possible to pick outsome of the deeper economic lows that are spaced 25 to 30 years apart(eg, 1868, 1894, 1921, 1946 and 1975). However, there are some deeplows that fell outside of the series (eg, 1908 and 1932).

What normally happens when faced with these apparent inconsis-tencies is that analysts assume that a) cycles of different lengths areindependent of one another and that b) the deeper the low, the longermust be the cycle that is operating. They would then go through thedata, searching for longer cycles centred on the Great Depression lowof 1932.

In fact, the depth of the cycle is not necessarily the most importantfactor in determining cycle periodicities. The important considerationis the intrinsic pattern. Deep lows may occur at the end of a cycle butthey can also occur at other points within a cycle. Deep lows within acycle can occur as a result of shocks, especially policy mistakes. Aswill be demonstrated, the Great Depression low did not occur becauseit coincided with the end of an important long-term cycle. It occurredbecause it was part of an energy gap within the economic system. Aswe saw in Chapter 14, an energy gap may more accurately define acycle high than a cycle low.

So the starting point in a search for cycles is twofold: either we canfind an obvious energy gap that can be used to locate an economicpeak and spread our analysis outwards around it; or we can look forthe centred three-up/three-down base cycle that is normally to befound at the beginning of a triad.

THE 1868–94 CYCLE

The first obvious candidate for an important base cycle is the 26-yearcycle that emerged after the end of the American Civil War. It fulfilsthe basic technical criterion of being a balanced six-wave (ie, three-up/three-down) cycle and it fulfils the economic criterion of being aperiod when adaptation to new circumstances was most likely. Themomentum pattern is shown in Figure 16.2, using two-yearpercentage rates of change on annual data. It consists of a three-waveup move (1–2–3) followed by a three-wave down move (A–B–C). Inline with the requirements of a base cycle, the pattern is balancedaround a centre point.

Recurrence in economic and financial activity 219

It is of some importance that this large adaptation cycle can easilybe divided into Juglar-length sub-cycles. The first one (1868–75)covers the period now known as the ‘Gilded Age’, as emotionsrecovered from the tragedy of the Civil War. The second period(1875–85) was the time when cultural deepening took place, as theNorth and South came a little closer together. And the third period(1885–94) was the period of cultural unease and disruption known asthe ‘Missionary Awakening’, characterized by labour violence andstudent evangelism.1

THE 1894–1921 CYCLE

The next cycle, in a sense, necessarily follows from this. After aperiod of rest and recuperation, creative energies re-emerge. In thiscase, a new generation ‘comes of age’. Figure 16.3 shows themomentum of US industrial production over the 27-year period from1894 to 1921. The period basically covers the huge adjustment in USsociety as it shifted to the mass production of industrial goods. As withmost innovation-related cycles, this one had a strong rightward biasand did not properly peak until 1916. Hence, the pattern consists of abase pattern (1–2); a trend wave (wave 3, albeit interrupted by a crisisin 1907); an energy gap (wave A) up to the First World War; a sharprecovery into a late peak (wave B); and a short, sharp downswing(wave C). This is the classic pattern for a trend cycle.


-15

-5

5

15

25

35

45

55

18

67

18

68

18

69

18

70

18

71

18

72

18

73

18

74

18

75

18

76

18

77

18

78

18

79

18

80

18

81

18

82

18

83

18

84

18

85

18

86

18

87

18

88

18

89

18

90

18

91

18

92

18

93

18

94

18

95

2-ye

ar %

ch

ang

e

1



2

B

A0

3

C

Figure 16.2 Post Civil War adaptation, 1868 to 1894

Like the cycle before it, this large cycle can be divided into threeseparate Juglar-length phases. The first phase (1894–1904) has thetechnical characteristics of a base cycle insofar as it has a balancedthree-up/three-down look to it. The second phase (1904–14) is theperiod most usually associated with mass production of consumergoods such as cars and telephones. The period was, however, inter-rupted by a financial crisis, which was precipitated by the failure ofthe Knickerbocker Trust in 1907. Nevertheless, output recovered. Thethird phase (1914–21) was dominated by the United States’involvement in the First World War. Note how the initial war-inducedsurge was followed by a sharp three-wave decline into the 1921 low.

THE 1921–46 CYCLE

An adaptation cycle and a regeneration cycle are inevitably followedby a crisis cycle. The 1921–46 cycle was the cycle that confirmed theaccuracy of this rule. Figure 16.4 shows the momentum of US indus-trial production over this period, measured by two-year percentagerates of change in the annual data. The chart embraces the Wall StreetCrash of 1929, the Great Depression of 1929–32 and the SecondWorld War of 1939–46. It is quite obvious that the pattern of the cycleconforms to the basic archetype that has already been described forterminal cycles: there is an early peak, and then a slow, but ultimatelydevastating, three-wave (A–B–C) fall into the final low.


Figure 16.3 Regeneration cycle, 1894 to 1921

-20

-10

0

10

20

30

40

18

94

18

95

18

96

18

97

18

98

18

99

19

00

19

01

19

02

19

03

19

04

19

05

19

06

19

07

19

08

19

09

19

10

19

11

19

12

19

13

19

14

19

15

19

16

19

17

19

18

19

19

19

20

19

21

19

22

2-ye

ar %

ch

ang

e

1



2

B

A

0

3

C

STRAUSS AND HOWE META-CYCLE

These three cycles are set out in sequence in Figure 16.5. Theiraverage periodicity is 26 years, which places them in the time spanappropriate to Berry cycles. The first such cycle, which covered the 26years of reconstruction after the Civil War, was thus an adaptationcycle. The second, which covered the 27 years of innovation in elec-tricity, electronics and cars, was a regeneration cycle. And the third,which covered the 25 years of disruption involving the GreatDepression and the Second World War, was a crisis cycle.

We thus have a set of three major Berry cycles that accord to thebasic blueprint – both in terms of pattern and time – that was set out inChapter 15. This, in turn, means that we have a triad of cycles thatconstitute some form of higher-level cycle. The time span of the threeBerry cycles, taken together, is 78 years. This is an acceptable timespan for a Strauss and Howe meta-cycle. We can thus conclude thatthe period that lasted from just after the end of the American Civil Warto the end of the Second World War was, indeed, a Strauss and Howemeta-cycle. It is no accident that the cycle was born out of a secularcrisis and ended in one.


-40

-30

-20

-10

0

10

20

30

40

50

60

70

19

21

19

22

19

23

19

24

19

25

19

26

19

27

19

28

19

29

19

30

19

31

19

32

19

33

19

34

19

35

19

36

19

37

19

38

19

39

19

40

19

41

19

42

19

43

19

44

19

45

19

46

19

47

2-ye

ar %

ch

ang

e

PeakUS Industrial Production

(Annual, monthly averages, 2-year % changes)Data sources: NBER & FRB

B

A C

Figure 16.4 Crisis, 1921 to 1946

THE BERRY TERMINAL CYCLE

Although the overall patterning is entirely consistent with the triadicmodel, it does raise a question about why 1946 is shown as the low ofthe terminal cycle, rather than the more traditional date of 1932. It isimportant to remember that the cycle is here being defined by thepatterns of group behaviour, not by specific events such as a very deeprecession low. Specific events are relevant – and will give a cycle itsessential flavour – but they do not define that cycle. So, we can gofurther and look at the shorter-term cycles within the larger-scale1921–46 cycle. Even a terminal cycle should exhibit signs of anunderlying triad of cycles. Figure 16.6 shows one-year percentagerates of change in the index of US industrial production between 1921and 1946. The three-wave-up/three-wave-down notation that definesthe locus of the whole cycle is shown in blocked alphanumerics(1–2–3/A–B–C) and the notation for the sub-cycles is shown in lower-case lettering and Roman numerals. It is quite clear that the 25-yearBerry crisis cycle that lasted from 1921 to 1946 consisted of threeJuglar-length sub-cycles.


-40

-20

0

20

40

60

1868

1871

1874

1877

1880

1883

1886

1889

1892

1895

1898

1901

1904

1907

1910

1913

1916

1919

1922

1925

1928

1931

1934

1937

1940

1943

1946

2-ye

ar %

cha

nge

US Industrial Production(Annual, averages of monthly data, 2-yr % changes)


1

B

A

3

2

1

C

B

A

3

2

Peak

C

B

A

Adaptation Regeneration Crisis

C

Figure 16.5 From the American Civil War to WWII

THE FIRST JUGLAR CYCLE

We can now see how this Berry crisis cycle breaks down into smaller-degree cycles. In theory, it should initially divide down into threecycles of approximately eight to nine years each. This would makethem Juglar cycles. The first such Juglar cycle lasted from March1921 to October 1930. It actually breaks down into a very clear six-wave pattern. In a terminal cycle, the momentum peak is very likely tooccur in the first sub-cycle. This is exactly what happened. Themomentum peak (at (i)) was high partly in response to the sharp low in1921. Nevertheless, subsequent peaks (at (iii) and (b)) were progres-sively lower. Hence, although the recuperation cycle contained a veryclear pattern of three waves up and three waves down, and althoughthere was an appropriate peak in the middle of the cycle in July 1925,the cycle contained warning biases.

THE SECOND JUGLAR CYCLE

The second cycle in the triad lasted from October 1930 to June 1938.It also breaks down into a very clear six-wave pattern. It is notable thatwave ii within this pattern made a new momentum low. This wasessentially because the authorities had made a mistake in allowing thecredit and monetary aggregates to contract. This policy mistake was a


-40

-20

0

20

40

60

80

Ma

r-2

1

Ma

r-2

2

Ma

r-2

3

Ma

r-2

4

Ma

r-2

5

Ma

r-2

6

Ma

r-2

7

Ma

r-2

8

Ma

r-2

9

Ma

r-3

0

Ma

r-3

1

Ma

r-3

2

Ma

r-3

3

Ma

r-3

4

Ma

r-3

5

Ma

r-3

6

Ma

r-3

7

Ma

r-3

8

Ma

r-3

9

Ma

r-4

0

Ma

r-4

1

Ma

r-4

2

Ma

r-4

3

Ma

r-4

4

Ma

r-4

5

Ma

r-4

6

1-ye

ar %

ch

ang

e 1

US Industrial Production(Monthly, 1-year % changes)


B

A

3

C

0

i

ii

b

a

iii

c2

(i)

(ii)

(b)

(a)

(iii)

(c)

Figure 16.6 The crisis triad, 1921 to 1946

shock that distorted the cycle; it was not an essential part of the cycle.After such a deep low, it would not have been surprising to see a sharprecovery. This latter duly occurred, assisted by Franklin D Roosevelt’s‘New Deal’ after he was elected in November 1932. Nevertheless, theeconomy needed more time to re-establish confidence. Productioncontracted again in 1937 and early 1938.

THE THIRD JUGLAR CYCLE

The third cycle lasted from May 1938 to February 1946. It reflectedthe impact of the war in Europe. Its position in the triad makes it theterminal sub-cycle within the bigger cycle. Initially, the United Stateswas not directly involved with the war. However, it became increas-ingly supportive of Britain’s efforts throughout 1940 and early 1941.This helped to provide a boost to the cycle, which therefore had anunusually centred look to it. Nevertheless, total output subsequentlycontracted, first as production facilities became focused on militaryhardware and then as the war in Europe finished in the spring of 1945.

CONFIRMING THE 1946 LOW

Measuring from 1868, the Strauss and Howe meta-cycle lasted for 78years. It contained three Berry cycles each averaging 26 years induration. The last cycle in the Berry triad – the 1921–46 crisis cycle –contained three clear Juglar sub-cycles, which took the economy intoits final low. These considerations are important. Had the final lowbeen placed in 1932, then the last phase would only have consisted oftwo cycles that had Juglar-type lengths. Further the time elapse ofthat last phase (1921–32) would have been only 17 per cent of theimplied Strauss and Howe meta-cycle (1868–1932), which is abouthalf the standard percentage. And the meta-cycle itself would havebeen unusually short at 64 years, which is about two-thirds of thestandard length.

The year 1946 therefore marked the end of an era, as measured by aStrauss and Howe meta-cycle. This is a critical conclusion, because itmeans that we can now start estimating where we are in the subse-quent meta-cycle. As a central estimate, the new era that started in1946 is destined to last 90 years. It could, of course, be about 10 yearslonger or shorter than this, depending on how the underlying Berry


and Juglar cycles evolve. Nevertheless, the new Strauss and Howecycle is scheduled to end somewhere between 2026 and 2046.

THE 1946–80 BERRY ADAPTATION CYCLE

The new era that began in 1946 started with a Berry adaptation cycle.We looked at this cycle briefly in Chapter 15. It was characterized bya large-scale government incursion into the economy and was ulti-mately inflationary. It lasted from 1946 to 1980. The time elapse ofthis cycle – at 34 years – was obviously longer than the Berry cyclesthat materialized in the previous meta-cycle. This is because theunderlying Juglar and Kitchin cycles were longer. It might be inter-esting to speculate why this might have been so but, as far as thisanalysis is concerned, it doesn’t strictly matter. All that we need toknow is where we are at any particular point in time.

Figure 16.7 shows the profile of US industrial production over theperiod 1946–80, measured in terms of two-year percentage rates ofchange. Data are, however, monthly, which allows for a little moreprecision. The chart covers the Korean War, the social revolution ofthe 1960s, the Vietnam War and the great inflation of the late 1970s.

This cycle was divided into three clear Juglar-length phases. Thefirst (February 1946 to April 1958) was a classic base cycle, with threewaves into the 1951 peak and then three waves down into the 1957


-40

-30

-20

-10

0

10

20

30

40

50

Jan-

46

Jan-

48

Jan-

50

Jan-

52

Jan-

54

Jan-

56

Jan-

58

Jan-

60

Jan-

62

Jan-

64

Jan-

66

Jan-

68

Jan-

70

Jan-

72

Jan-

74

Jan-

76

Jan-

78

Jan-

80

2-ye

ar %

cha

nge

1

1 3

A

3

2C

B

B

A

2C

Peak

C

B

A



Recuperation Innovation Disruption

Figure 16.7 The Berry adaptation cycle, 1946 to 1980

low. It covered the basic post-war reconstruction of industry andbusiness. The second cycle in the triad (April 1958 to November1970) spanned the period known as the ‘Swinging Sixties’. It markedthe transition in the economy to mass consumption. It was thereforean innovation cycle and – not surprisingly – was biased to the right.The last cycle (November 1970 to August 1980) was the cycle ofdisruption, related to an explosion in the prices both of consumergoods and of raw materials, especially oil. This cycle followed theclassic pattern that has already been outlined – namely, an early peak,followed by a three-wave fall into a final low.

THE 1970–80 JUGLAR CYCLE

The last cycle needs a little explanation. It has traditionally beenassumed that 1975 marked the major cycle low of the period. Demandhad been badly hit in 1974 by a combination of higher interest ratesand the sharp rise in oil prices after the 1973 Yom Kippur War.However, on the basis of the model that is being presented here, thefinal low of the crisis phase of the post-war adaptation cycle could nothave been in 1975. For a start, it would have meant that the third cyclein the triad would have only been four years long, which is less than15 per cent of what would then have been the whole cycle (1946–75).Second, inflationary expectations were not at that time broughtproperly under control because governments still did not see thatpublic sector deficit financing was at the root of the problem. Finally,there was a second oil price shock after the Ayatollah’s Revolution inIran, which deflated the US economy into a second low. Such an eventis typical of a terminal cycle. The point is that the 1975 low wascreated – as was the low in 1932 – by a shock to the system. It was notgenerated by the process that creates a normal cycle low. Hence thecorrect placement of the low is in 1980. This ensures that the thirdcycle in the post-1946 triad was 10 years in length. This is of the sameorder of magnitude as the previous two.

THE KITCHIN TRIAD

A closer inspection of the 1970–80 Juglar cycle, using a shortermomentum index, reveals that it sub-divides into three Kitchin cycles.Figure 16.8 shows the annualized three-month percentage rates of


change in US industrial production between November 1970 and June1980. This is a 10-year cycle that breaks down into three sub-cycles,each lasting about 3.2 years, or 38 months. Moreover, the pattern ofthe sub-cycles contains the expected three-up/three-down profile.

The first Kitchin cycle is very clear. It peaked halfway through itseventual time elapse and finished in February 1974. The second cyclestarted on time, but was deflated by the impact of the huge increase inoil prices that occurred in December 1973. Wave 2 was thereforevery deep and made a new low. As mentioned in Chapter 9, this is byno means an unusual phenomenon. Nevertheless, there was a wave 3recovery into the summer of 1975. The subsequent wave A was quitemild, allowing wave B to peak at high levels. This gave a rightwardbias to the middle wave. Finally, the third Kitchin cycle looksremarkably like the terminal Juglar cycle just described. It has asharp initial rally and then a slower three-wave fall. It is also shorterthan its predecessors. And note this: the pattern in Figure 16.8 isalmost identical to the pattern in Figure 16.3.

Hence, the patterns of the Kitchin cycles reproduced the patterns ofthe Juglar cycles, which repeated the patterns of the Berry cycles,which mirrored the patterns of the Strauss and Howe meta-cycle.Cycles nest in fractal-like hierarchies.


-35

-25

-15

-5

5

15

25

35

Nov

-70

Mar

-71

Jul-7

1

Nov

-71

Mar

-72

Jul-7

2

Nov

-72

Mar

-73

Jul-7

3

Nov

-73

Mar

-74

Jul-7

4

Nov

-74

Mar

-75

Jul-7

5

Nov

-75

Mar

-76

Jul-7

6

Nov

-76

Mar

-77

Jul-7

7

Nov

-77

Mar

-78

Jul-7

8

Nov

-78

Mar

-79

Jul-7

9

Nov

-79

Mar

-80

Jul-8

0

Ann

ualiz

ed 3

-mon

th %

cha

nge 1

1

3

A

3

2

B B

A

2

Peak

B

A

Recovery RecessionAdvance

US Industrial Production(Monthly, annualized 3-month % change)


C

C

C

Figure 16.8 US industrial production, November 1970 to June 1980

THE POST-1980 BERRYREGENERATION CYCLE

The model suggests that, after a period of adaptation, the economyenters a period of genuine change, involving qualitative evolution inthe physical and intellectual infrastructure. The Berry cycle thatstarted in 1980 is such a cycle; and, on the basis of the average ofBerry cycles since 1868, it is scheduled to finish in 2010. There issome evidence that the cycle has lengthened since the end of theSecond World War, which means that it could finish as late as 2014.Normally, this spread of end dates would cause forecasters to turn toother methods. The point, however, is that we will likely be able totrack the cycle into its final low just by focusing on the pattern.

THE JUGLAR RECUPERATION ANDINNOVATION CYCLES

Shown in Figure 16.9 is the two-year percentage change in themonthly index of US industrial production between 1980 and 2005. Itcovers the initial recovery out of the 1980 low, the deep recession ofthe early 1990s, the long subsequent boom associated with theinfotech revolution, and the ‘unexpected’ collapse of 2000–01.


Figure 16.9 The Berry regeneration cycle, 1980 to 2005

-10

-5

0

5

10

15

Feb-

80

Feb-

81

Feb-

82

Feb-

83

Feb-

84

Feb-

85

Feb-

86

Feb-

87

Feb-

88

Feb-

89

Feb-

90

Feb-

91

Feb-

92

Feb-

93

Feb-

94

Feb-

95

Feb-

96

Feb-

97

Feb-

98

Feb-

99

Feb-

00

Feb-

01

Feb-

02

Feb-

03

Feb-

04

Feb-

05

Feb-

06

1-ye

ar %

cha

nge

1

A

3

0

i

ii

b

a

iii

c2

(i)

(ii)

(b)

(a)

(iii)

(c)

US Industrial Production(Monthly, 1-year % changes)


Up to the end of 2001, therefore, the Berry regeneration cycle hadspanned two of its three Juglar cycles. The first of these (July 1980 toMarch 1991) was a classic base cycle in its own right. It had a clearsix-wave pattern consisting of three waves up ((i), (ii), (iii)) and threewaves down ((a), (b), (c)). As occasionally happens, wave (ii) in thispattern made a new low. This was partly caused by the fact thatinterest rates were not only high, but also rose again in the early part of1982. So there was a ‘second dip’ to the recession. However, theoverall cycle was basically centred and, when completed, it formedwaves 1 and 2 of the higher-level Berry cycle.

The second pattern (March 1991 to November 2001) was the inno-vation cycle proper. It incorporated the spread of personal computers,the adoption of Internet usage and the conversion to mobile tele-phones. The cycle was biased to the right and the recession, when itcame, was sudden and seemingly unexpected.

KITCHIN CYCLES

If we look a little more closely at this Juglar innovation cycle, we cansee just how closely it conforms to the archetypal triadic patterning fora complete cycle. Shown in Figure 16.10 is the three-month annu-alized percentage change in the monthly index of US industrialproduction from January 1991 to November 2001. The starting date ofthe cycle is slightly earlier because of the shorter momentum indicatorbut it demonstrates that the Juglar cycle breaks down into threesmaller Kitchin cycles.

The first one (January 1991 to April 1995) was a standard base, orrecovery, cycle as the economy recovered from the traumaticrecession of 1990. It consisted of the archetypal centred six-wavepattern associated with such a cycle. The second cycle (April 1995 toJuly 1998) was the Kitchin advance cycle. It had a long upward biasand therefore peaked late. The third cycle (July 1998 to November2001) was the Kitchin recession cycle. It peaked early and then had along three-wave correction into the final low.

CONCLUSION

The vibrations in industrial production since the end of the AmericanCivil War confirm that an economic cycle, of whatever duration,


evolves according to a basic pattern. The pattern consists of three sub-cycles. The first cycle is a base (or detachment) cycle. It consists of abalanced six-wave pattern and peaks in the middle. The second cycleis a trend (or transition) cycle that incorporates the social and techno-logical changes of the period. It, too, consists of six waves. However,it is biased to the right and tends to peak late. The third cycle is aterminal (or detachment) cycle. The underlying pattern consists of sixwaves, but the occurrence of shocks tends to suppress some of them.In particular, the first downturn within the early part of the cycle canact as an energy gap, or shock, that reverses the polarity of the system.The cycle therefore tends to present as a sharp rise into an early peak,followed by an extended three-wave fall into a final low.

The basic pattern has been found to exist within a series of cyclesthat harmonize with each other. It exists in a 90-year Strauss andHowe meta-cycle, a 30-year Berry infrastructure cycle, a 10-yearJuglar investment cycle and a 3.33-year Kitchin cycle.

These are extraordinary conclusions. They mean that economiccycles are both ordered and predictable.

NOTE

1. See Strauss, William and Howe, Neil (1991) Generations: The history ofAmerica’s future, 1584 to 2069, William Morrow, New York.


-15

-10

-5

0

5

10

15

Ja

n-9

1

Ju

l-9

1

Ja

n-9

2

Ju

l-9

2

Ja

n-9

3

Ju

l-9

3

Ja

n-9

4

Ju

l-9

4

Ja

n-9

5

Ju

l-9

5

Ja

n-9

6

Ju

l-9

6

Ja

n-9

7

Ju

l-9

7

Ja

n-9

8

Ju

l-9

8

Ja

n-9

9

Ju

l-9

9

Ja

n-0

0

Ju

l-0

0

Ja

n-0

1

Ju

l-0

1

3-m

on

th %

pa

1 B

A

3

2

1 B

A

3

2

Peak

C

B

A

C

C

US Industrial Production(Monthly, 3-month % p.a.)


Ja

n-0

2

Figure 16.10 The Juglar innovation cycle, 1991 to 2001

232

17

Integrating the cycles

INTRODUCTION

The inference from the foregoing analysis is that group activity resultsin clearly identifiable patterns of behaviour. The patterns are derivedfrom a single archetypal pattern, evolve in groups of three andoscillate rhythmically. This phenomenon, which is currently unrecog-nized by the social sciences, hints at the existence of a profound depthof order to social, political and economic development. In otherwords, although evolution is always a creative response to changingcircumstances, the mechanisms by which that evolution takes placeare not random.

If this is so, then we have the possibility of comparing currentdevelopments (whatever they may be and whenever they occur) withprevious episodes of history, in order to see what can be learnt. This issomething that historians have always attempted to do. The differencehere is that we have important criteria for ensuring that we arecomparing like with like.

HISTORICAL SCHEMATIC

The schematic diagram in Figure 17.1 incorporates the broad charac-teristics that we have been able to identify for each of the Berry andJuglar cycles that have evolved since the beginning of the 19thcentury. The cycles are measured from low points in economic

Integrating the cycles 233

momentum, using annual data. Since we are here looking at broadtrends and essential characteristics of cycles, the use of annual datadoes not matter too much. However, the dating varies slightly whenshorter-term momentum indicators are used. For example, in the1930s, the US economy almost certainly moved from the adaptationstage to the regeneration stage of the Berry cycle in 1931. However,the regeneration stage was initially biased downwards such that indus-trial production continued to fall into 1932. Inevitably, therefore, thediagram must be used with appropriate caution.

With this qualification in mind, the diagram incorporates the basicthree-phase model of cyclical behaviour that was shown in Figure15.2, in Chapter 15. Here, each of the three phases is a Juglar cycle;and three Juglar cycles constitute a Berry cycle. Each Juglar cycle willbe either a base cycle, a trend cycle or a terminal cycle. Each Berrycycle will be either a base cycle, a trend cycle or a terminal cycle.These terms will, of course, have a different quality of meaning,depending on the level of the hierarchy being considered.

Shown in the lower part of the diagram are the dates of each of themajor Juglar cycle turning points that have already been identified.These turning points have been identified by the pattern of the cyclerather than by the periodicity of the cycle.

Between each pair of dates are shown the essential characteristics ofthe Juglar cycle in question. These characteristics are historically iden-tifiable and will, in essence, involve recuperation, innovation ordisruption. Similarly, down the right-hand side of the lower part of thediagram are the indications as to whether a specific Berry cycle is

Figure 17.1 Characteristics of Berry and Juglar cycles

Stylized

Terrorism and WMD

orientated to adaptation, regeneration or crisis. We can thus locate theessential ‘flavour’of any particular time period from its position on thediagram.

SURVEY OF THE DIAGRAM

What is so striking about the diagram is the natural way in which thevarious distinctions can be drawn and applied. At the Berry level, it is,for example, quite obvious that the 1868 to 1894 period should havebeen an adaptation period following the Civil War. And it is quiteobvious that the 1921 to 1946 cycle was a huge crisis cycle. Theprimary characteristics of the Berry cycles have therefore occurredprecisely where they should have done.

Meanwhile, at the Juglar level, history is quite clear that theimportant periods of very active change in the last century were: thedecade prior to the First World War (mass production); the SwingingSixties (mass consumption); and, more recently, the infotech revo-lution (information technology). Again, periods of recognizableeconomic and social change occur precisely where they should. Onevery obvious inference is that periods of important adjustment occurevery 30 years or so.

Having thus locked the diagram in place, other periods can easily befitted in. For example, we can see how precisely the three Juglar cyclesthat evolved between 1868 and 1894 fitted in. First, the period 1868 to1875 is now known as the ‘Gilded Age’, reflecting the relief that theCivil War was over. Then 1875 to 1885 saw something of a culturalrevolution as the North and South started to integrate. Finally, theMissionary Awakening that started in 1885 embraced a period ofunrest. Indeed, it was a complete contrast to the quiet deepening of theprevious 10 years. There were violent labour disputes and there was anexplosion of missionary activity. Both, in their way, were aconfrontation with the existing order of things: the labour disputeswere very much related to a desire for a radical change in the domesticculture; the missionary activity was energized by a desire to change theexternal world. It seems that a polarization of attitudes and variousforms of unrest are part and parcel of any Juglar disruption cycle.

On this analysis, it is not therefore surprising to find that everyJuglar disruption cycle in Figure 17.1 has been dominated by an inten-sification of aggression. Hence, the Civil War impacted during the1858–68 Juglar disruption cycle; the First World War occurred in the1914–21 cycle; the Second World War impacted during the 1938–46


cycle; and the US invasion of Cambodia and the the peak of the ColdWar (as well as the Yom Kippur War) occurred during the 1970–80cycle. It is, therefore, consistent with the model that the War onTerrorism has dominated the start of the terminal Juglar cycle withinthe current Berry trend cycle.

Where world war has dominated a disruption cycle, the room fordomestic political confrontation has been reduced. Nevertheless, thepotential for such confrontation is never very far away. The experienceof the Missionary Awakening period has already been mentioned.There was also labour unrest in the 1970s. In 1978 in the United States,for example, the miners went on strike and newspapers were shut downfor months. In the UK (which follows the US cycle quite closely), thewinter of 1978/79 became known as the ‘Winter of Discontent’.

Naturally enough, the traumas of a disruption cycle generate a needfor recuperation. Hence the 10-year period following each Juglardisruption cycle has been very much associated with a sense of relief,especially in the first five years or so. The ‘Gilded Age’ has alreadybeen mentioned. The recovery from the First World War was asso-ciated with the ‘Roaring Twenties’; the recovery from the SecondWorld War was very much a period of reconstruction and generatedcultural phenomena such as the ‘New Look’, ‘Teddy Boys’ and rockand roll. And the recovery from the war-peppered and inflationary1970s was marked by a decisive shift towards conservative politics(Ronald Reagan in the United States and Margaret Thatcher in the UK)and the start of the lengthy campaign against large government deficitsand rapid monetary growth.

Finally, it needs to be noted that the last phase of the recuperationcycle is usually marked by a deep recession. The need to detach fromthe old order means that some of its entrenched attitudes have to beeradicated. Some of these attitudes are often the outmoded methods ofdealing with a crisis in the first place. The Great Depression of1929–32, the recession of 1957–58 and the slump of 1990–91occurred during the last phase of a Juglar recuperation cycle.

USING THE DIAGRAM

One definite inference from this summary is that the diagram can beused to assess the tone of current social and economic developments.There may be some uncertainty about precise start and end dates, andthere is always the danger of unforeseen shocks. Nevertheless, it ispossible to embed forecasts for the economy and financial markets in


specific assumptions that will almost certainly be justified. Forexample, it is possible to say that the Juglar cycle that began in 2001 willbe characterized by persistent international conflict and labour unrest.

In addition, we can assume that fluctuations in industrialproduction during this Juglar cycle will exhibit a very clear pattern.Specifically, there will be an initial recovery, which will be followedby a slow three-wave fall into a late period low. This implies that therewill be at least two very difficult periods, of which the second one(into the cycle low) will probably be the worst. What we cannot yetknow is how strong the intermediate recovery will be.

This is where the bigger picture may provide some help. Anotherway of using the diagram is to see what is unlikely over any particularperiod. In the 1990s, for example, we could validly have regardedinternational tension as being a low risk. Indeed, we might haveconcluded this even before the fall of the Berlin Wall in 1989. This didnot mean that warfare wouldn’t happen during the 1990s, just that itwas unlikely to be a long-term dominant factor as far as the domesticeconomy and mass public attention were concerned. Hence, the GulfWar, which impacted at the end of the 1980–91 recuperation cycle,was quick and (from the West’s point of view) effective. Oil prices fellback after the war and the global economy quickly recovered. Thiscontrasts, for example, with the intra-cycle war events of 1970 to1980, which resulted in a rapid expansion of money growth and anirreversible explosion in oil prices. The US economy (and othereconomies) suffered from these events.

What is unlikely in the current Juglar cycle is another 1930s-stylerecession. For a start, the current Berry cycle is a regeneration cycle,not a crisis cycle; and the Great Depression occurred in a Berry crisiscycle. Second, the tone of the current Juglar cycle is simply not thesame as that which gripped the US economy during the Depressionyears. This needs some explanation.

THE GREAT DEPRESSION

As indicated in Chapter 16, socio-economic behaviour expresses itselfin a different blend of affective, somatic and cognitive characteristicsin each Juglar cycle. A Juglar disruption cycle is, of course, dominatedby fear. Ultimately, however, fear evokes a response from those whoare closely involved with it. The most obvious response is frominvestors and employers: the former sell stock and the latter lay offlabour. However, the response that in many ways is the most


important (in relation, that is, to the onset of an economic depression)comes from the authorities. They, too, will be driven by fear. But thequality of their policy response will be influenced by the somatic andcognitive milieu of the time.

In a Juglar disruption cycle, the dominant somatic reaction is one ofresistance. Initially this takes the form of resistance to any more change(ie, satiation with innovation). Subsequently, however, it takes theform of resistance to losing the gains that were achieved during theinnovation cycle. Meanwhile, the dominant cognitive reaction is one ofaccusation. Initially, this takes the form of passively blaming externalevents for changing circumstances. Subsequently, however, it can takethe form of direct action against a perceived aggressor. This is one ofthe reasons why the prolonged economic difficulties of a Juglardisruption cycle are associated with persistent civil unrest and/ormilitary action.

However, the combination of fear, resistance and accusation can bea powerful stimulus for governments to take action on the economy.Governments become interventionist. Monetary and fiscal policybecomes very pro-active. Tariff barriers are erected against foreigncompetition. Failing industries are supported by direct intervention.This action cannot bypass the economic contraction entirely but it canameliorate that contraction.

This is in direct contrast to the matrix of indolence, need for recu-peration, and loss of self-assertiveness that can bleed into businessactivity and policy-making during a recuperation cycle. Normally,this may not matter very much; it is just part of the ebb and flow oflife. However, under particularly trying conditions – such as in a Berrycrisis cycle – it can matter a great deal. The problem is that no one isable to take control once things start to go wrong. A sort of lazinessand fatalistic acceptance settles over business and politics. Mistakesget made and a genuine psychological depression sets in. This is oneof the aspects of the Great Depression that is least commented on – itwas, above all else, a psychological phenomenon.

THE NEXT BERRY CRISIS CYCLE

One inference, therefore, is that the 10-year terminal cycle that startedin 2001 will be difficult. Specifically, it seems likely to drive the economy from above-trend growth to below-trend growth.However, the pattern clearly confirms that it is not destined to have the


psychological overtones that characterize a major economicdepression. In fact, if any such problem were to occur, it would mostlikely impact during the next Juglar cycle. There are two reasons forthis. First, the next 30-year Berry cycle will probably be a crisis cycle.Second, the next Juglar cycle, which will also start at the same time,will probably be a recuperation cycle. The implication is that thedownturn of this next Juglar cycle could bring with it a depression-liketrauma and thereby trigger a crisis period that could last until 2040.

The general themes for the 30-year period are shown in Figure 17.2.Taking our lead from the experience of the 1920s and 1930s, the middleof the next decade (ie, 2015 or thereabouts) will probably mark the endof a prolonged period of economic growth, and a satiated economy willdrop into an energy gap. If, then, the patterns of history continue torepeat themselves, 2020–30 will see the beginnings of a revolutionaryupheaval in attitudes towards politics and economic policy, and2030–40 will be blemished by global conflict. We shall see.

KONDRATYEV WAVES

This ‘forecast’ raises the question of the relationship between thecurrent cyclical pattern and the Kondratyev price cycle, which wementioned in Chapter 15. Specifically, if a genuine crisis is going to


Stylized

Terrorism and WMD

Figure 17.2 The Berry crisis cycle, 2010 to 2040

start sometime in the next decade, what might commodity prices bedoing? Figure 17.3 shows the one-year percentage change in the annualclosing level of the three-year average of US wholesale commodityprices. The time series starts from the late 18th century; so it is a rela-tively long series. What is apparent is that there is a basic oscillation,measured from low to low, that takes about 28 years. However, thereare a number of other distinctive features. The first is that the cyclestend to be biased to the right, so that highs occur very late in the cycleand are followed by a sharp drop. The second is that the most dominanthighs are associated with major wars. And the third feature is that eachpair of oscillations is very biased upwards to the right, such that the firstoscillation forms a base and the second oscillation involves a three-phase impulse wave and a final collapse. Moreover, the peaks are 28years apart, so a double oscillation lasts for about 56 years. This is theaccepted periodicity of the Kondratyev wave.

So, there certainly seems to be some ordered process at work.However, the important point is that the highs are almost certainlypreludes to energy gaps, which are reflected in the sharpness of thesubsequent drop. The highs are thus genuine cycle highs, but the lowsare not necessarily always proper cycle lows. As was demonstrated inFigure 14.7, lows may be caused by shocks to the system. On thisbasis, the correct version of Figure 17.3 is as shown in Figure 17.4.The periodicity is still 56 years, but the lows are now in differentplaces. Specifically, the lows are in 1842, 1895 and 1954.


Figure 17.3 US wholesale price inflation, 1796 to 2001

-20

-10

0

10

20

30

40

50

60

1796

1804

1812

1820

1828

1836

1844

1852

1860

1868

1876

1884

1892

1900

1908

1916

1924

1932

1940

1948

1956

1964

1972

1980

1988

1996

2004

2012

2020

1-ye

ar %

cha

nge

NapoleonicWar

American Civil War

First World War

Vietnam WarMiddle East War

Second World War

US Producer Prices: All Commodities(Annual close, 1-year % change of three year average)

Data sources: Global Financial Data & Federal Reserve Board

One important inference is that we can use these timings to identifyperiods of inflation and periods of disinflation. That is, the 28-yearperiod prior to a Kondratyev peak is inflationary and the 28-year periodafter that peak is disinflationary.

However, we can take this analysis an important step further. Oneof the important ideas that we are pursuing is that historical patternsreproduce themselves over the course of comparable cycles. Figure17.5 takes the 1841 to 1895 cycle shown in Figure 17.4 and compares


-20

-10

0

10

20

30

40

50

60

1796

1804

1812

1820

1828

1836

1844

1852

1860

1868

1876

1884

1892

1900

1908

1916

1924

1932

1940

1948

1956

1964

1972

1980

1988

1996

2004

2012

2020

1-ye

ar %

cha

nge

NapoleonicWar

American Civil War First World

WarVietnam War

Middle East War

Second World War

US Producer Prices: All Commodities(Annual close, 1-year % change of three year average)

Data source: Global Financial Data & Federal Reserve Board

Figure 17.4 The Kondratyev cycle in US wholesale prices

Figure 17.5 Comparison of cycles, 1841–95 and 1895–1954

-20

-10

0

10

20

30

40

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58

Years since low (1841 to 1895)

Perc

enta

ge c

hang

e

-20

-10

0

10

20

30

400 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58


Perc

enta

ge c

hang

e

1841 to 1895(l.h.scale, lower time axis)

1895 to 1954(r.h.scale, upper time axis)

US Producer Prices(Annual, 1-year % change)


it with the subsequent cycle that ran from 1895 to 1954. There is atruly remarkable correlation between the two patterns. In particular,there is a major cycle peak about 22 years into the cycle. Whatemerges, however, is that the later cycle was probably longer than itspredecessor, by up to five years. Taking the average of the two cycles,the cycle peaks seem to occur 39 per cent into the whole cycle.

Next, we can compare the 1895 to 1954 cycle with the most recentcycle, which evolved after 1954. This time, however, it is necessary toadjust the patterns, by plotting each cycle on separate axes, so that themajor turning points coincide (see Figure 17.6). Once this is done,there is another remarkable correlation between the patterns,providing it is allowed that the cycle that started in 1954 either endedin 2003 or is destined to have another sharp drop very shortly.Actually, when the most recent cycle is compared with the 1841 to1895 cycle, it looks reasonably clear that the former has indeed ended.At the comparable stage of the 1841 to 1895 cycle, prices had alreadystarted to rise sharply (see Figure 17.7). This suggests that the timeelapse of the most recent cycle is 49 years, rather than 56 years.Importantly, this would also mean that the cycle peak in year 20 wouldhave occurred 40 per cent of the way through it, which is not signifi-cantly different to the timings for its predecessors.

If we now translate the general information from the above analysisinto the terms of Figure 17.1, something very interesting emerges (seeFigure 17.8). First, the main Kondratyev price lows (eg, 1895 and1954) occur during Juglar recuperation cycles. Second, Kondratyev



-20

-10

0

10

20

30

40

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58


Perc

enta

ge c

hang

e

-10

-5

0

5

10

15

200 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52


Perc

enta

ge c

hang

e



US Wholesale Prices(Annual, 1-year % change)


price highs (eg, 1864, 1917, 1974) occur during Juglar disruptioncycles. Moreover, these highs are war-related. However, the precisetiming of the price highs and price lows within the appropriate Juglarcycle seems not to be a constant. In other words, there is a slightmismatch between prices and output.1

What is clear is that Kondratyev price highs occur during Juglardisruption cycles and that such Juglar cycles are likely to involveserious international tension. What is also clear is that the Juglar recu-peration cycle after a major price peak is particularly dangerous.There were deep recessions in 1873–75, 1929–32 and 1990–91.


-20

-10

0

10

20

30

40

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58


Perc

enta

ge c

hang

e

-10

-5

0

5

10

15

200 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52


Perc

enta

ge c

hang

e



US Producer Prices(Annual, 1-year % change)



Figure 17.8 The Kondratyev price cycle

KONDRATYEV WAVES AND STRAUSS ANDHOWE META-CYCLES

There are now a number of additional points that can be made aboutthe diagrams shown in Figures 17.1 and 17.8, which actually help toconfirm their validity. The first point is that the Crash of 1873, theWall Street Crash of 1929 and the Equity Crash of 1987 all occurred inthe downwave of Juglar recuperation cycles. Further, they all occurredafter a major Kondratyev price high. This is compelling evidence thatthe model is an accurate one.

Nevertheless, there are differences. The 1873 event occurred in aBerry adaptation cycle, and Congress limited the circulation of notesand coins. The 1929 episode occurred in a Berry crisis cycle, and theauthorities made the mistake of allowing the money and credit aggre-gates to contract. The 1987 event occurred in a Berry regenerationcycle, and the authorities reacted with a monetary stimulus.

There is a second point about the diagrams, which is in many waysmore important. The diagram quite clearly indicates that the GreatDepression is not specifically a Kondratyev phenomenon. Indeed, aspresented, the Wall Street Crash and Depression were part of thebeginning of a crisis period, rather than part of the end of one. This isentirely consistent with the idea offered by Strauss and Howe thatmajor socio-economic crises occur every 90 years or so. In the contextof the basic cycle mechanism, the Crash/Depression presents itself asan energy gap in the system. In the context of the 90-year Strauss andHowe meta-cycle, such an energy gap is likely to occur in the thirdBerry cycle (ie, the Berry crisis cycle) in a triad of such cycles.Specifically, it is likely to occur in the first downturn of the first Juglarcycle in that Berry cycle. Measuring from 1868 in Figure 17.1, thedownturn could have been expected in the late 1920s – which isprecisely where it occurred.

Finally, there is the interesting conundrum of whether or not theKondratyev wave that began in 1954 has actually finished or not. Upto 1954, the periodicity of the Kondratyev wave has actually beensteadier than that of the underlying economic cycles, with a perio-dicity of 56 years. On the basis of simple mathematics, the next lowafter 1954 is scheduled for 2010; and we have already indicated thatan output low is scheduled for 2010. However, a more detailedanalysis of the last Kondratyev cycle suggests that it may already bemoving into another upwave. This not only supports the view that thelong wave is not a mechanical phenomenon, but also suggests that theeconomic low scheduled for around 2010 may not actually be severe.


There are two good reasons to suspect that this may be the case. Thefirst is that the authorities are, in general terms, aware of thephenomenon of end-cycle deflation. They have moved strongly toavoid it. The second reason is that the economic confidence modelshown in Figure 17.2 reveals that any end-cycle economic weaknesswill not be part of an energy gap. This latter is scheduled for 2015 (orthereabouts). Such a gap will necessarily arise out of excesses, and it isreasonable to assume that those excesses will be sourced in theattempts to avoid recession during the 2001–10 period. What is more,whether the low of the most recent Kondratyev cycle occurred in 2003or is likely to impact in 2010, one thing seems certain: the next energygap will be associated with inflation, not deflation.

CONCLUSION

Forecasting financial markets will probably never be a perfect science;but it can be a less stressful art. In this respect, the suggestion thateconomic activity exhibits patterned rhythms is an important asset. If ananalyst can identify the current location of the economy in the context ofan evolving pattern, then there is a good chance that it will be possible todistinguish a little more clearly between temporary fluctuations andimportant trends.

This chapter has therefore surveyed in some detail the evolution ofeconomic cycles dating back to the end of the American Civil War. It hasfound that certain economic phenomena reproduce themselves periodi-cally. Quite specifically, it has been found that the economy enters aperiod of innovation, or change, every 30 years or so, as the energy of anew mix of generations starts to make itself felt. This period of inno-vation is preceded by a period of gradual awakening and is succeeded bya period of retrenchment. Each of these phases lasts about 10 years.

The presence of this phenomenon is highly suggestive of an orderedprocess at work. Moreover, it confirms that economic theory’s assumptionof rational behaviour by independent individuals is deeply flawed.

NOTE1. Professor Berry assigns two ‘Berry cycles’ to one beat of the Kondratyev wave.

This means that each Kondratyev price low and price high is accompanied by arecession. The former could be a depression; the latter is a stagflation. SeeBerry, Brian J L (1991) Long-Wave Rhythms in Economic Development andPoliticial Behavior, Johns Hopkins, Baltimore.


245

18

Forecasting with cycles

INTRODUCTION

In the last chapter, we revealed the presence of order in the evolutionof economic cycles. Importantly, this order is based on logic, and theevidence supports the theory. We thus have the capability of under-standing the economic context within which financial market cyclesoccur, without having to resort to a detailed analysis of daily andmonthly statistics. We can know, for example, when we are in a crisiscycle that is vulnerable to shocks, especially war-related shocks. Wecan know whether this cycle is likely to be inflationary or not. We canjudge the likely pattern of the cycle over time. And we can use historyto estimate when the important turning points are going to be.

This is a huge step forward. Nevertheless, the periodicity of a cyclewill still vary through time. No matter how careful we are, the timeelapse of a cycle can expand or contract and throw us off balance. Thecritical point, however, is that, despite the potential variation in theperiodicity, the basic pattern of the cycle remains unchanged. The art,therefore, is to track the pattern of a cycle in real time.

One of the ways of doing this is to use a previous cycle as a blueprintand measure the evolution of the contemporaneous cycle against it.

THE CURRENT BERRY ECONOMIC CYCLE

Shown in Figure 18.1 is the two-year rate of change in the monthlyindex of US industrial production over two time periods. The first

time period is the whole Berry adaptation cycle that evolved between1946 and 1980. The second is the incomplete Berry regeneration cyclethat started in 1980. The latter Berry cycle has so far negotiated twoJuglar cycles – namely 1980 to 1991, and 1991 to 2001 – and is now inits third.

The secret of making the comparison is to track the evolution of thecurrent cycle on a different time axis to the original cycle. This meansthat any variations in the time elapse of the cycle can be absorbed,while holding the pattern constant. In Figure 18.1, the 1946–80 cycle– now to be our blueprint – is tracked along the lower time axis and thesubsequent Berry cycle that started in 1980 is tracked along the uppertime axis.

CYCLE ALIGNMENT

The first task is to align the three-up/three-down pattern of the1980–91 Juglar cycle with the three-up/three-down pattern of the1946–57 Juglar cycle. This should be relatively easy to do becauseboth cycles are base cycles and should therefore have the same clearshape. This shape normally incorporates a mid-cycle peak.1 Once thealignment has been done, the 1991–2001 Juglar cycle then easilycoincides with the 1957–70 cycle.


-40

-30

-20

-10

0

10

20

30

40

Feb-

46

Feb-

48

Feb-

50

Feb-

52

Feb-

54

Feb-

56

Feb-

58

Feb-

60

Feb-

62

Feb-

64

Feb-

66

Feb-

68

Feb-

70

Feb-

72

Feb-

74

Feb-

76

Feb-

78

Feb-

80

2-ye

ar %

cha

nge

-20

-15

-10

-5

0

5

10

15

20

Jul-8

0

Jul-8

2

Jul-8

4

Jul-8

6

Jul-8

8

Jul-9

0

Jul-9

2

Jul-9

4

Jul-9

6

Jul-9

8

Jul-0

0

Jul-0

2

Jul-0

4

Jul-0

6

Jul-0

8

Jul-1

0

2-ye

ar %

cha

nge

Jul 80 to date(r.h.scale, upper time axis)

Feb 46 to Jul 80(l.h.scale, lower time axis)


Data source: Federal Reserve Board

Figure 18.1 Juglar economic cycles compared

This alignment seems amazing, although it is actually quite normal.It means that it was possible to confirm both the 1991 low and the2001 low, as they were occurring.

The second task is to estimate the likely duration of the whole Berrycycle that started in 1980. Looking at the top horizontal axis, it can beseen that, in order to make the momentum troughs on the chartcoincide, the total time elapse of the post-1980 cycle would have to be355 months, or 29.6 years. Since the cycle started in July 1980, thismeans that the expected low for the whole cycle is early 2010. This isa clear confirmation of the date given in the previous chapter, whichwas calculated using averages of previous cycle periodicities.

There is a new piece of information here. This is that the post-1980cycle is likely to be shorter than its 1946–80 counterpart. The lattertook 413 months (or 34.4 years) to complete. This confirms the valueof looking at the pattern and, as a corollary, the danger of concen-trating on periodicities.

SENTIMENT IN A JUGLARINNOVATION CYCLE

Although it is more than adequate to work just with economicstatistics, it is often helpful to track sentiment indicators as well. Theseare usually released more quickly than official data and, indeed, mayactually inflect before the official data. One of the primary indices touse in relation to the US economy is the composite index of producerexpectations that is produced by the Institute of Supply Management(ISM). Comparable indices are produced by the Confederation ofBritish Industry (CBI) in the UK and by the Institute for EconomicResearch (IFO) in Germany.

Such indices are very sensitive. So they are particularly useful forshort-term work and for comparing cycles directly with one another.Figure 18.2 compares the Juglar innovation cycle in the ISM indexbetween January 1958 and November 1970 with the Juglar innovationcycle in that same index between February 1991 and November 2001.The ISM index has been re-based to 100 at the cycle start dates.

What is very clear is that, once the two completed cycles are alignedwith one another, their basic shapes are very similar. This again meansthat it was possible to recognize in late 2001 that an end-cycleinflexion was occurring.

Forecasting with cycles 247

Moreover, it was possible to do so despite the inherent difference inthe time elapse between the two cycles in Figure 18.2. This isimportant. The original 1958–70 cycle lasted for 154 months but the1991–2001 cycle took only 129 months. This was a difference of morethan two years. In terms of history, such a difference is not unusual.However, it could have been a major problem for an analyst who wasfollowing fixed periodicities.

It was therefore possible to conclude in late 2001 that the USeconomy was transiting from a Juglar innovation cycle into a Juglardisruption cycle, with all the problems that such a transition wouldentail.

SENTIMENT IN THE JUGLAR CRISIS CYCLEHowever, even crisis cycles necessarily start with a recovery of somesort. The Juglar disruption cycle that started in late 2001 was nodifferent. However, the move from an innovation cycle into a crisiscycle involves a move from supply-led growth to demand-led growth.That is, capital investment subsides, although inventory investmentrecovers. So the analyst can always expect the initial upswing of anyJuglar disruption cycle to be more quantitative than qualitative. Thisdistinction is very important. It takes most people a very long time torecognize that things have actually changed and that the dynamism ofthe innovation cycle cannot immediately be resumed.


100

120

140

160

180

200

220

240Ja

n-58

Jul-5

8

Jan-

59

Jul-5

9

Jan-

60

Jul-6

0

Jan-

61

Jul-6

1

Jan-

62

Jul-6

2

Jan-

63

Jul-6

3

Jan-

64

Jul-6

4

Jan-

65

Jul-6

5

Jan-

66

Jul-6

6

Jan-

67

Jul-6

7

Jan-

68

Jul-6

8

Jan-

69

Jul-6

9

Jan-

70

Jul-7

0

Jan-

71

Inde

x le

vel

100

110

120

130

140

150

160

170Fe

b-91

Jun-

91

Oct

-91

Feb-

92Ju

n-92

Oct

-92

Feb-

93Ju

n-93

Oct

-93

Feb-

94Ju

n-94

Oct

-94

Feb-

95

Jun-

95O

ct-9

5Fe

b-96

Jun-

96

Oct

-96

Feb-

97Ju

n-97

Oct

-97

Feb-

98Ju

n-98

Oct

-98

Feb-

99

Jun-

99O

ct-9

9Fe

b-00

Jun-

00O

ct-0

0Fe

b-01

Jun-

01

Oct

-01

Feb-

02

Inde

x le

vel

Jan 91 to date(r.h.scale, upper time axis)

Jan 58–Jan 73(l.h.scale, lower time

axis)

ISM Composite Index(Rebased, cycle start =100)

Data source: Institute of Supply Management

Figure 18.2 Juglar innovation cycles in the ISM index

An analyst could then estimate the likely strength and duration ofthe post-2001 recovery by comparing it with the one that occurredafter the 1970 low. The first thing to do, therefore, would be to extendthe locus of the older cycle. The new cycle should continue to move inparallel with it. This is exactly what happened (see Figure 18.3).

COMPARISONS WITH PREVIOUS CYCLES

The next stage, quite obviously, would be to look at the whole crisiscycle, rather than just the first stage. Figure 18.4 shows the movementin the ISM index of producer expectations between October 2001 andOctober 2005 in the context of the same index between November1970 and May 1980.

As already argued, the pattern of the 1970–80 Juglar disruptioncycle in Figure 18.4 is archetypal. It shows the initial rally into a peakand then a long three-wave decline. Because of the danger of unex-pected events, there is always the possibility of divergences from thenorm, especially in the form of a shorter-than-expected cycle.Nevertheless, as of October 2002, an analyst could reasonably haveconcluded that the 2001–10 Juglar disruption cycle was evolvingaccording to precedent. Furthermore, an analyst could also have


100

120

140

160

180

200

220

240

Jan-

58Ju

l-58

Jan-

59Ju

l-59

Jan-

60Ju

l-60

Jan-

61Ju

l-61

Jan-

62Ju

l-62

Jan-

63Ju

l-63

Jan-

64Ju

l-64

Jan-

65Ju

l-65

Jan-

66Ju

l-66

Jan-

67Ju

l-67

Jan-

68Ju

l-68

Jan-

69Ju

l-69

Jan-

70Ju

l-70

Jan-

71Ju

l-71

Jan-

72Ju

l-72

Jan-

73Ju

l-73

Jan-

74Ju

l-74

Inde

x le

vel

100

110

120

130

140

150

160

170

Feb-

91

Aug

-91

Feb-

92

Aug

-92

Feb-

93

Aug

-93

Feb-

94

Aug

-94

Feb-

95

Aug

-95

Feb-

96

Aug

-96

Feb-

97

Aug

-97

Feb-

98

Aug

-98

Feb-

99

Aug

-99

Feb-

00

Aug

-00

Feb-

01

Aug

-01

Feb-

02

Aug

-02

Feb-

03

Aug

-03

Feb-

04

Aug

-04

Feb-

05

Inde

x le

vel

Jan 91 to date(r.h.scale, upper time axis)

Jan 58–Jan 73(l.h.scale, lower time

axis)

ISM Composite Index(Rebased, cycle start =100)


Figure 18.3 Extending into the next cycle

concluded that there was still some time left before the next seriousdownleg would impact. In Figure 18.4, the real problems in the1970–80 cycle started where shown, in early 1974.

In the event, although activity did indeed slow during 2005, it wasnot a rerun of 1974 and, by October 2005, the ISM index had started torecover. The important point, however, is that the pattern was exactlyas predicted. Just to confirm the importance of this pattern, we cancompare the 1970–80 fluctuations in the ISM index with anotherperiod of serious dislocation. Figure 18.5 accordingly compares theISM index during the 1970–80 Juglar disruption cycle with move-ments in industrial production during the 1921–46 Berry crisis cycle.The Juglar cycle peaked later in the context of the whole cycle thandid the Berry cycle, but the patterns were very similar.

It can clearly be seen that, despite their different time spans, theserious problems that were specific to each cycle impacted at almostexactly the same time location within the cycle. This is shown in Figure18.5 by the dotted vertical line. Within the 1921–46 Berry cycle, theequity crash happened in 1929; within the 1970–80 Juglar cycle, theeconomic recession started in early 1974. Both periods of weaknessstarted to impact 33 per cent into the total time elapse of the cycle. Thiscalculation is, of course, made in retrospect. Nevertheless, it impliesthat the cycle takes the shape that it does because the economy fails torecover when it should do – namely, about one-third of the way into thecycle. In other words, the low of the first sub-cycle fails to provide


25

30

35

40

45

50

55

60

65

70

75

Nov

-70

Mar

-71

Jul-7

1

Nov

-71

Mar

-72

Jul-7

2

Nov

-72

Mar

-73

Jul-7

3

Nov

-73

Mar

-74

Jul-7

4

Nov

-74

Mar

-75

Jul-7

5

Nov

-75

Mar

-76

Jul-7

6

Nov

-76

Mar

-77

Jul-7

7

Nov

-77

Mar

-78

Jul-7

8

Nov

-78

Mar

-79

Jul-7

9

Nov

-79

Mar

-80

ISM

inde

x le

vel

25

30

35

40

45

50

55

60

65

70

75

Oct

-01

Jan-

02

Apr

-02

Jul-0

2O

ct-0

2Ja

n-03

Apr

-03

Jul-0

3O

ct-0

3Ja

n-04

Apr

-04

Jul-0

4O

ct-0

4Ja

n-05

Apr

-05

Jul-0

5O

ct-0

5Ja

n-06

Apr

-06

Jul-0

6O

ct-0

6Ja

n-07

Apr

-07

Jul-0

7O

ct-0

7Ja

n-08

Apr

-08

Jul-0

8O

ct-0

8Ja

n-09

Apr

-09

Jul-0

9O

ct-0

9Ja

n-10

ISM

inde

x le

vel

Nov 70 to May 80(lower time axis)

Oct 01 to Oct 05(upper time axis)

US ISM Composite Index(Monthly, seasonally adjusted)


Star

t of p

robl

em

--

Figure 18.4 The Juglar disruption cycle

support and the recovery that should initiate the second sub-cycleaccordingly fails to materialize on time. This phenomenon is one of thecharacteristics of an energy gap. As at the end of 2005, such a gap hadnot impacted the US economy. It remains to be seen how closely the restof the 2001–10(?) Juglar cycle will follow the pattern.

THE KITCHIN CYCLE

The idea that we can anticipate, and then track, a new cycle as itdevelops is exciting in itself. However, we can fine-tune the fore-casting process by delving down into the shorter-term cycles. Forexample, in the context of a Juglar cycle, we can look at the 3.33-yearKitchin cycles that initiate it. This is particularly appropriate insofaras Kitchin cycles are primarily concerned with inventory adjustments.It is almost certainly the case that a Juglar cycle will end with arundown in inventories and that the new one will therefore be trig-gered by a recovery in inventories. If the basic model is correct thenthe first Kitchin cycle in any Juglar cycle should consist of an arche-typal three-up/three-down pattern.

Shown in Figure 18.6 is the profile of the ISM producer confidenceindex, measured in terms of three-month actual changes in that index,for the first 24 to 32 months of the Juglar cycles that began in 1970,


25

30

35

40

45

50

55

60

65

70

75N

ov-7

0

Mar

-71

Jul-7

1

Nov

-71

Mar

-72

Jul-7

2

Nov

-72

Mar

-73

Jul-7

3

Nov

-73

Mar

-74

Jul-7

4

Nov

-74

Mar

-75

Jul-7

5

Nov

-75

Mar

-76

Jul-7

6

Nov

-76

Mar

-77

Jul-7

7

Nov

-77

Mar

-78

Jul-7

8

Nov

-78

Mar

-79

Jul-7

9

Nov

-79

Mar

-80

ISM

com

posi

te in

dex

-50

-40

-30

-20

-10

0

10

20

30

40

50

60

70D

ec-2

1

Dec

-22

Dec

-23

Dec

-24

Dec

-25

Dec

-26

Dec

-27

Dec

-28

Dec

-29

Dec

-30

Dec

-31

Dec

-32

Dec

-33

Dec

-34

Dec

-35

Dec

-36

Dec

-37

Dec

-38

Dec

-39

Dec

-40

Dec

-41

Dec

-42

Dec

-43

Dec

-44

Dec

-45

Dec

-46

2-ye

ar %

cha

nge

in in

dust

rial p

rodu

ctio

n

ISM composite indexMonthly, Nov 70 to May 80(l.h.scale, lower time axis)

US industrial productionAnnual, 1921–1946

(r.h.scale, upper time axis)

Peak

A

B

CStar

t of p

robl

em

Figure 18.5 Juglar disruption cycle compared with Berry crisis cycle

1980 and 1991. The oscillations in the momentum indices are marked0–1–2–3–A–B–C. Hence the first sub-cycle covers 0–1–2, the secondsub-cycle is 2–3–A and the third is A–B–C. It is quite clear that,although there is some divergence in the timing and extent ofmomentum in the third sub-cycle, the patterns accord to the basicmodel. If we now lay the Kitchin cycle that began in late 2001 on thematrix of previous cycles, we can see how closely it tracked them up toJune 2004 (see Figure 18.7).

ESTIMATING THE TIMING OFTURNING POINTS

This process again suggests that we can use the timing of previousturning points to forecast future ones. This can be done using just oneprevious cycle, or it can be done using an average of previous cycles.Suppose that previous cycles mirrored the basic model, and registeredimportant lows at 33 per cent and 67 per cent of their total time elapse,and registered important highs at 17 per cent, 50 per cent and 83 percent of their total time elapse. It would be reasonable to expect that thecurrent evolving cycle would also register the same pattern.

The unknown, however, is precisely how long the current cycle isactually going to take. It could be shorter or longer than the average,even though the turning points emerge in the ‘correct’ place. This


-10

-5

0

5

10

15

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Months since cycle low

3 m

th a

ctua

l cha

nge

in in

dex

-30

-20

-10

0

10

20

300 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28


3 m

th a

ctua

l cha

nge

in in

dex

Nov 70 to Jul 73(l.h.scale) Jan 91 to Apr 93

(l.h.scale)

ISM Composite Index(Monthly, seasonally adjusted)


(r.h.scale)

3

1

C

B

A20

Figure 18.6 The Kitchin cycle

May 80to

May 82

problem can be dealt with as soon as turning points have started tomaterialize. If we know that the first peak is going to occur 17 per centinto the cycle then, once we have an inflexion point that looks appro-priate, we can start to calculate the eventual likely outcome. That is,

Possible cycle length = time from start of cycle to first peak / 0.17

For example, if we were tracking a Juglar cycle, we might expect thefirst high to arrive after about 20 months. Suppose that such a highthen occurred after 22 months, the estimated cycle length would be22 / 0.17 = 129.4 months. In other words, the estimated cycle lengthafter the first peak would be 10.78 years.

From this, we could then start estimating the other intra-cycleturning points. We would expect the first important low to occur one-third of the way into the cycle. That is,

First low = estimated total length of cycle / 0.33

Hence, for a 10.78-year cycle, the first low would be calculated tooccur after (10.78 / 0.33 =) 3.56 years, which is 42.7 months.

Then we could calculate the likely timing of the second high (ie, themid-point) of the cycle. That is,

Second high = estimated cycle length / 0.50

So, for a 10.78-year cycle, the second high would be estimated tooccur after (10.78 / 0.5 =) 5.39 years, which is 64.7 months.


Figure 18.7 The Kitchin cycle

-10

-5

0

5

10

15

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32


3 m

th a

ctua

l cha

nge

in in

dex

-30

-20

-10

0

10

20

300 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28


3 m

th a

ctua

l cha

nge

in in

dex

Nov 70 to Jul 73(l.h.scale) Jan 91 to Apr 93

(l.h.scale)

ISM Composite Index(Monthly, seasonally adjusted)


May 1980 to May 82(r.h.scale)

3

1

C

B

A2

Oct 01 to Jun 04(l.h.scale)

0

CUMULATIVE ESTIMATIONWe are thus starting to get quite a lot of information, just from onepoint of reference. Nevertheless, we know that, however close theactual outcome might be to the basic model, there are bound to besignificant variations. The analysis, therefore, is cumulative – that is, ittakes the variations into account as they occur. Hence, suppose that thefirst low of the Juglar cycle arrives on time (ie, after 3.56 years), butthat the second high arrives, not after 5.39 years, but after 5.08 years(ie, 62 months). This would indicate that the cycle length might nowbe shorter than previously estimated. That is, since

Possible cycle length = time from start of cycle to second peak / 0.5

then calculating the cycle just from the second peak would indicate alength of (5.08 / 0.5 =) 10.16 years.

However, the implications of the first and second turning points stillhold good, so the best solution is to average the calculations. Hence,the new central estimate for the total cycle length after three turningpoints becomes the average prediction from those three turning points.In our example, therefore, the new central prediction after two highsand an interim low is (10.78 + 10.78 + 10.16) / 3 = 10.57 years. This is,indeed, shorter than the first estimates. And we can now use the newestimated length to calculate future turning points.

It follows from this that the closer we get to the end of the cycle, themore definite we will be about the actual point at which it will end. Giventhat the psychological atmosphere at the last stage of any cycle is likely tobe one of capitulation, such information may be absolutely priceless.

THE LEAD FROM FINANCIAL MARKETS TOTHE ECONOMY

Importantly, the capitulation normally takes place in the equity marketabout four to six months before it occurs in manufacturers’ stock levels.Hence, the closer that we get to the end of a large economic cycle, themore important it becomes to track the behaviour of the equity market asan item of information. If equities start to rally strongly, within sixmonths of a projected low in output, then the chances are that the equitymarket is right. The very fact that equities might have broken loose fromfeedback with the economy is a valuable piece of information in its ownright. The more important point, however, is that the rally will beconfirming the imminence of a cyclical turn in output.


Shown in Table 18.1 is the relationship between a turn in the absolutelevel in the Dow and turning points in the level of production at thestart of each of the major Juglar cycles since 1957. They confirm thatthe average equity lead time is about five months.

CONCLUSION

This chapter has demonstrated that, once cycles have been properlycategorized, they can be used in a very powerful manner. Specifically,the pattern of pre-existing cycles can be used to track the pattern of anewly emerging one, and those patterns can be used to look into thefuture. The simplest method is to compare similar cycles to one another.Hence we can compare base, trend and terminal cycles with other base,trend and terminal cycles. However, we can also compare differentcycles with one another. Hence we can compare a base cycle with atrend cycle or a terminal cycle. Or we can compare a Berry cycle with aJuglar cycle. Or we can even compare financial market cycles withoutput cycles. This can be particularly fruitful if a major economic lowis expected. The point is that the basic model of economic and financialbehaviour that was revealed in Chapter 16 is universal: it applies to allgroup behaviour, at all levels of the hierarchy.

NOTE1. The 1980–91 cycle made a new low in 1982. This might have created some

uncertainty about the correct date for the end of the previous crisis cycle. Infact, the validity of the relationship shown in Figure 18.1 can be confirmed byusing the index of producer sentiment, calculated by the Institute of SupplyManagement. This index did not make a new low in 1982.


Table 18.1 Equity lead times over output

DJIA low (level) US output low (level) Equity lead (months)

Dec 57 Apr 58 4Jun 70 Nov 70 5Jul 82 Dec 82 5Oct 90 Mar 91 6Sep 01 Dec 01 3

Average: 4.6

256

19

Finding cycles: a casestudy

INTRODUCTION

There is, of course, no point in attempting to use cycles, if you can’tfind them in the first place. In fact, finding them is something of an art.Sometimes they leap off the chart at you. At other times, however,they remain tantalizingly elusive. It is like looking at a distant shipthrough a powerful telescope: if the focus is too close, you see only thesea; if the focus is too distant, you see only the sky. You need thecorrect focus to give you the desired result. With cycles, the lens isfocused by adjusting the speed of the momentum indicator.

GUIDELINES

There are no hard-and-fast rules for finding cycles. Nevertheless,there are certain initial guidelines that can be used. The first point toremember is that it is the pattern that defines the cycle, not theprecision of the periodicity (although the periodicity is relevant). Sothe starting point is to look for a recognizable three-up/three-downpattern. The second point to remember is that cycles operate in triads.So, if there is one easily recognizable pattern, there ought to be twoothers associated with it. The third point to remember is that verysharp drops could be energy gaps, which define a peak rather than a

Finding cycles: a case study 257

trough (see Chapter 14). So, unusual lows may not actually mark theend of a cycle.

In addition, there is a broader point. This is that the cycles are nestedin hierarchical layers. Consequently, and providing the data set beinganalysed is sufficiently detailed, it is very helpful to look for threelayers of cycles – namely, the immediate level, the one above it and theone below it.

PRACTICAL PROCESS

The practical process then starts by calculating a long-term picture of amarket or economic indicator by using a long-term momentum indi-cator. One very useful indicator is a two-year percentage rate ofchange. This has the dual advantage of smoothing out some of thevolatility in a series and providing a close estimate of the timing ofmost of the peaks and troughs. Sometimes, this will be sufficient,particularly with very slow-moving processes like the economy itself.Often, however, it will only provide a broad indication of where themajor turning points are, and a further refocusing has to be done.

UK GILTS

Shown in Figure 19.1 is a two-year percentage rate of change of the UKgovernment bond market (‘gilts’), represented by the Treasury 2½ percent bond, redeemable on or after 1 April 1975. As already mentioned,this bond is usually called an ‘irredeemable’ because it is unlikely to beredeemed until redemption yields fall below 2½ per cent. It has the bigadvantage for analysis of not being influenced by the so-called ‘pull toredemption’, whereby the price of a bond tends to be drawn towards itsredemption price (usually 100) as its maturity shortens.

Shown on the same chart is the actual price performance of the bondbetween 1971 and late 2002. It therefore covers the tail end of theinflationary bear market of the 1970s, and the great disinflationary bullmarket of the 1980s and 1990s.

The first thing to notice in the chart is that, between October 1974and January 1996, there was a huge three-up/three-down momentumpattern, marked 1–2–3/A–B–C. This was reflected in a similarlyprofiled price oscillation, with a slight upward bias, which started inDecember 1974 and ended in April 1996. This, in effect, was the

massive pattern that formed the base for the subsequent impulse wavethat occurred between April 1996 and January 1999. The base patternhas the requisite archetypal formation and took 19.26 years. So it is agood starting point.

The earlier analysis of triads in Chapter 17 suggests that the 19.26-year cycle ought to sub-divide into three approximately equal cyclebeats of (19.26/3 =) 7.09 years. When this calculation is applied to theactual lows in bond prices, starting at December 1974, it does seem topick out important turning points. The theoretical lows of a potential7.09-year cycle are shown by the upward-pointing arrows in Figure19.1. Obviously the start and end dates coincide with the arrows.However, the end dates of the sub-cycles do not. The first sub-cycleshould have ended in January 1982 but it ended in September 1981.The second sub-cycle should have ended in February 1989 but itended in April 1990. The first discrepancy is not necessarily signif-icant in the context of a 7.09-year cycle; however, the seconddiscrepancy is significant and will need an explanation. In themeantime, if we project this periodicity forward, out of the basepattern, the estimated date of the next low would be April 2003.

CALCULATING PEAKSThe next step is to analyse the peaks. One point to remember is thatmomentum peaks are often biased in relation both to the centre point


-60

-40

-20

0

20

40

60

80

Jan

-71

Jan

-72

Jan

-73

Jan

-74

Jan

-75

Jan

-76

Jan

-77

Jan

-78

Jan

-79

Jan

-80

Jan

-81

Jan

-82

Jan

-83

Jan

-84

Jan

-85

Jan

-86

Jan

-87

Jan

-88

Jan

-89

Jan

-90

Jan

-91

Jan

-92

Jan

-93

Jan

-94

Jan

-95

Jan

-96

Jan

-97

Jan

-98

Jan

-99

Jan

-00

Jan

-01

Jan

-02

Jan

-03

Jan

-04

2-ye

ar%

ch

ang

e

10

15

20

25

30

35

40

45

50

55

60

Pric

e

UK Treasury 2½ % 'Irredeemable '(Monthly close and 2-year % change)


Price(r.h.scale)

Momentum

(l.h.scale)

1

2

3

A

B

C

Figure 19.1 The UK gilt market, 1971 to 2002

of a cycle and in relation to the actual price peak. Momentum willoften accelerate coming out of a cycle low and will often be deceler-ating at a cycle high. This helps to validate the cycle (see Chapter 13)but it is also a reminder that we might be looking for ‘non-confirmed’price highs at around the middle of the cycle.

The obvious price peaks are as shown by the downward-pointingarrows in Figure 19.1. They occurred in December 1977, April 1986and December 1993. Importantly, the second of these peaks is ‘non-confirmed’ by the two-year momentum, which indicates its impor-tance. The average period is – quite clearly – eight years. Further,projecting forward gives a potential peak in January 2001. There was,indeed, a peak at that time. However, there was an even moreimportant one in January 1999, which was two years earlier. So, thereis only a loose match between the periodicities of the highs and lows.

REFOCUSING MOMENTUM

The next step, therefore, is to recalibrate the momentum indicatorover shorter time periods. The most obvious one to use is a one-yearpercentage rate of change. This is shown in Figure 19.2, together withthe actual level of prices.


-40

-30

-20

-10

0

10

20

30

40

50

60

Ja

n-7

1

Ja

n-7

2

Ja

n-7

3

Ja

n-7

4

Ja

n-7

5

Ja

n-7

6

Ja

n-7

7

Ja

n-7

8

Ja

n-7

9

Ja

n-8

0

Ja

n-8

1

Ja

n-8

2

Ja

n-8

3

Ja

n-8

4

Ja

n-8

5

Ja

n-8

6

Ja

n-8

7

Ja

n-8

8

Ja

n-8

9

Ja

n-9

0

Ja

n-9

1

Ja

n-9

2

Ja

n-9

3

Ja

n-9

4

Ja

n-9

5

Ja

n-9

6

Ja

n-9

7

Ja

n-9

8

Ja

n-9

9

Ja

n-0

0

Ja

n-0

1

Ja

n-0

2

Ja

n-0

3

Ja

n-0

4

1-ye

ar %

ch

ang

e

10

15

20

25

30

35

40

45

50

55

60

Pric

e

UK Treasury 2½% 'Irredeemable'(Monthly close and 1-year % change)


Price(r.h.scale)

Momentum

(l.h.scale)

1

2

3

A

B

C

Figure 19.2 The UK gilt market, 1971 to 2002

Now, a slightly different picture starts to emerge. First, if we lookfor the three-up/three-down pattern, the first cycle in the sequencemay have ended in February 1980 rather than May 1981. This isshown by the 1–2–3–A–B–C pattern on the chart. Second, theDecember 1993 to June 1994 fall in prices (the bond ‘crash’) looksvery much like an energy gap. In this case, the momentum low thatwas registered in December 1994 was not a cycle low, but the low inApril 1996 probably was. This latter coincides with the conclusionfrom the two-year momentum index. Moreover, we now also have aclearer view of the importance of the peak in late 1993.

If we now assume that February 1980 and April 1996 are validcycle lows, then they can only be matched if the cycle periodicity is5.34 years. This is shown by the upward-pointing arrows in Figure19.2. The low in April 1990 now comes much closer to being properlyincluded in the cycle, even though the theoretical timing for it wasDecember 1990.

At this stage of the analysis, therefore, we have the possibility of thepresence of either a 7.09- to 8-year cycle or a 5.34-year cycle.Interestingly, the 5.34-year cycle is just over 75 per cent of the 7.09-year cycle, which means that there is some form of harmonic rela-tionship working. If the two cycles were separate, four 5.34-yearcycles would coincide with three 7.09-year cycles. That is, they wouldapproximately coincide every 19.25 years or so.

THE IMPORTANCE OF THREE

There are two potential solutions here. The first is to assume that thetwo cycles are independent of one another. The second is to look forreasons why one of the two cycles might be false. One of the cluescomes from the fact that three longer cycles equal four shorter cycles.Specifically, what might be happening is that shocks and energy gapsin a genuine 5.34-year cycle may have created conditions that couldbe misread as lows of a false 7.09-year cycle. Figure 19.3 shows howthis might occur. The upward-pointing arrows show the 5.34-yearcycle lows. The downward-pointing arrows show the 7.09-year cyclelows. One of the issues is whether the low of April 1990 belongs to a5.34-year cycle or a 7.09-year cycle. A 5.34-year cycle would divideinto three sub-cycles of about 1.78 years (ie, just over 21 months); and5.34 + 1.78 = 7.12. In other words, the 7.09-year cycle could be theshorter 5.34-year cycle with an extra 1.78-year cycle added on. In thiscase, unexpectedly deep lows in the 5.34-year cycle could be misread


as lows in a 7.09-year cycle. Conversely, weak lows in the 5.34-yearcycle might be ignored altogether.

This may be exactly what has happened. It remains true that theprice low in September 1981 was deeper, and that the low in March1985 was weaker, than might have been expected for a 5.34-yearcycle. However, the April 1990 low can be seen as the early end of the5.34-year cycle rather than the very late end of a 7.09-year cycle.

SCANNING FOR TRIADS

The secret, obviously, is to look for triads, rather than just for periodicrhythms. So we can now take this a step further. We have already iden-tified the 5.34-year cycle that started in December 1974 as being anarchetypal three-up/three-down cycle. See Figure 19.2 again. Further,it ‘looks’ centred. This means that it could be a base cycle within thecontext of a batch of three. This makes sense inasmuch as it emergedafter the traumatic bear market of 1973/74. According to theschematic model outlined in Chapter 17, the next cycle is likely to bea trend cycle, which would have a rightward bias. A glance at Figure19.2 shows that the cycle from February 1980 to March 1985 doesindeed have an upward bias, such that the March 1985 low hardlyregisters. What is also clear, however, is that the first 1.78-year sub-cycle creates a very deep low before the trend starts. Let us, therefore,tighten the focus a little further by looking at six-month percentagerates of change (see Figure 19.4).

This chart takes the same form as its predecessors. Now, however,we can start being clearer about the influence of the cycle groupings.The first beat of the proposed 5.34-year cycle (November 1974 to


Figure 19.3 Confusion between cycles

March 1980) is marked on the chart as ‘First’. The second beat (March1980 to June 1985) is marked ‘Second’. Then there is a third beat ofthe cycle (June 1985 to April 1990), marked ‘Third’. These three beatsof the 5.34-year cycle, which in total take 15 to 16 years to evolve,form the big base pattern to which we have already referred. This ismarked ‘Base’. It is actually the cycle that ushered in the removal ofinflationary psychology, so it is also a ‘detachment’ cycle.

In the triadic model, this base pattern should lead into a big trendcycle, also lasting 15 to 16 years. This latter is not scheduled to enduntil late 2006/early 2007. However, it can be seen from the chart thatit is undoubtedly a trend cycle. This cycle is marked ‘Trend’ in Figure19.4 and it is the cycle that is involved with the move to genuinelylower inflationary expectations. So it is a ‘transition’ cycle.

The final 15- to 16-year cycle still lies well ahead. It should start inlate 2006/early 2007 and end in 2022/2023. It will be the cycle thatfully embeds disinflation in the economy. Unfortunately, it is likely tohave to do so in the context of falling bond prices.

THE FIRST CYCLE WITHIN THE BASE CYCLE

The first 5.34-year cycle within the base cycle (which, for simplicity,we will assume has a periodicity of 16 years) has a very clear arche-


-40

-30

-20

-10

0

10

20

30

40

50

Jun-6

9

Jun-7

0

Jun-7

1

Jun-7

2

Jun-7

3

Jun-7

4

Jun-7

5

Jun-7

6

Jun-7

7

Jun-7

8

Jun-7

9

Jun-8

0

Jun-8

1

Jun-8

2

Jun-8

3

Jun-8

4

Jun-8

5

Jun-8

6

Jun-8

7

Jun-8

8

Jun-8

9

Jun-9

0

Jun-9

1

Jun-9

2

Jun-9

3

Jun-9

4

Jun-9

5

Jun-9

6

Jun-9

7

Jun-9

8

Jun-9

9

Jun-0

0

Jun-0

1

Jun-0

2

6-m

on

th %

0

10

20

30

40

50

60

Pric

e le

vel

UK Treasury 2½% 'Irredeemable'(Monthly close and 6-month % change. Source: Datastream)

First Second Third First Second

Base Trend

Momentum

(l.h.scale)

Price level(r.h.scale)

Figure 19.4 Using six-month rates of change

typal three-up/three-down pattern in both price and momentum (seeFigure 19.5). The theoretical sub-cycle lows, calculated as one-thirdand two-thirds respectively of the final cycle length, are indicated bythe upward-pointing arrows. The first sub-cycle ends more or less ontime. The second sub-cycle ends somewhat late but neverthelessallows a rally prior to the final fall.

THE SECOND CYCLE

The second beat of the 5.34-year cycle has a clear three-up/three-down pattern in momentum terms but not in price terms (see Figure19.6). Prices basically trend upwards over the period, which confirmsthat the second cycle is a trend cycle within the overarching basecycle. The theoretical sub-cycle lows, calculated as one-third and two-thirds of the actual cycle, are shown as upward-pointing arrows. Asalready mentioned, the first such sub-cycle of this second 5.34-yearcycle makes a new low in both price and momentum terms. This is notunusual. The point is that a trend cannot fully develop until themarket, or economic index, has properly bottomed. This is particu-larly true of ‘nested’ base cycles – that is, lower-level base cycleswithin higher-level base cycles. The 1.78-year cycle was a base cyclewithin the 5.34-year trend cycle but was obviously still subject to thepull of the 5.34-year base cycle from which it was born.


-20

-15

-10

-5

0

5

10

15

20

25

30

Aug-7

4

Oct-

74

De

c-7

4

Fe

b-7

5

Apr-

75

Jun-7

5

Aug-7

5

Oct-

75

De

c-7

5

Fe

b-7

6

Apr-

76

Jun-7

6

Aug-7

6

Oct-

76

De

c-7

6

Fe

b-7

7

Apr-

77

Jun-7

7

Aug-7

7

Oct-

77

De

c-7

7

Fe

b-7

8

Apr-

78

Jun-7

8

Aug-7

8

Oct-

78

De

c-7

8

Fe

b-7

9

Apr-

79

Jun-7

9

Aug-7

9

Oct-

79

De

c-7

9

Fe

b-8

0

6-m

on

th %

ch

ang

e

13

15

17

19

21

23

25

Pric

e

UK Treasury 2½% 'Irredeemable'(Monthly close and 6-month % change)


Price(r.h.scale)

Momentum

(l.h.scale)

1

2

3

A

B

C

Figure 19.5 The first 5.34-year cycle

THE THIRD CYCLE

The third beat of the proposed 5.34-year cycle has a much morevolatile look than either of its predecessors (see Figure 19.7). This isbecause it is a terminal cycle in its own right. It is therefore subject toan air of crisis and a vulnerability to shocks. The theoretical lows ofthe sub-cycles are indicated by upward-pointing arrows. Theyactually coincide with near-term lows. However, the overarchingprofile is an early peak (in April 1986) followed by a slow three-wave(A–B–C) fall into a relatively deep low. The peak occurs, as is usual,in the first sub-cycle. The A-wave fall is an energy gap, and istherefore quite sharp. The B-wave is a long rally, on progressivelyweaker momentum. The C-wave is a deep fall into the final low. It isa classic pattern.

THE 16-YEAR TREND CYCLE

After the completion of the third (terminal) 5.34-year cycle in the 16-year base cycle, a new 16-year cycle starts. This is a trend cycle and,as such, should be characterized by a strong uptrend for part of itspattern. As has already been pointed out, this is exactly whathappened (see Figure 19.4 again). The period April 1990 to late 2002


-20

-10

0

10

20

30

40

Fe

b-8

0

Apr-

80

Jun-8

0

Aug-8

0

Oct-

80

De

c-8

0

Fe

b-8

1

Apr-

81

Jun-8

1

Aug-8

1

Oct-

81

De

c-8

1

Fe

b-8

2

Apr-

82

Jun-8

2

Aug-8

2

Oct-

82

De

c-8

2

Fe

b-8

3

Apr-

83

Jun-8

3

Aug-8

3

Oct-

83

De

c-8

3

Fe

b-8

4

Apr-

84

Jun-8

4

Aug-8

4

Oct-

84

De

c-8

4

Fe

b-8

5

6-m

on

th %

ch

ang

e

16

18

20

22

24

26

28

Pric

e



Price(r.h.scale)

Momentum

(l.h.scale)

1

2

3

A

B

C

Figure 19.6 The second 5.34-year cycle

was dominated by a strong uptrend in bond prices. In principle, thetrend cycle should consist of three 5.34-year cycles. Two of theseappear to have completed. The first, which lasted from April 1990 toApril 1996, had something of an upward bias but was in essence stillthe base cycle within the higher-order trend cycle. The second cycle,which lasted from April 1996 until (probably) June 2001, was thetrend cycle proper. Prices moved sharply higher during this cycle. Thethird cycle (the terminal cycle) is scheduled to finish between April2006 (which is 16.02 years – ie, 3 × 5.34 years – after April 1990) andOctober 2006 (which is 5.34 years after June 2001).

THE FIRST CYCLE WITHIN THE TREND CYCLE

The first (base) cycle within the 16-year trend cycle actually lasted sixyears, from April 1990 to April 1996. It was therefore longer than theaverage (see Figure 19.8). The theoretical timings of the sub-cyclelows, based on the completion of the whole base cycle, is shown bythe upward-pointing arrows. The pattern of the first sub-cycle (whichlasted about two years) is not clear from the behaviour of prices them-selves. The pattern does, however, reveal itself in the associated six-month rate of change.

The pattern of the second sub-cycle was dominated by an acceler-ation of prices into a peak in December 1993. The peak was relatively


Figure 19.7 The third 5.34-year cycle

-25

-20

-15

-10

-5

0

5

10

15

20

Fe

b-8

5

Apr-

85

Jun-8

5

Aug-8

5

Oct-

85

De

c-8

5

Fe

b-8

6

Apr-

86

Jun-8

6

Aug-8

6

Oct-

86

De

c-8

6

Fe

b-8

7

Apr-

87

Jun-8

7

Aug-8

7

Oct-

87

De

c-8

7

Fe

b-8

8

Apr-

88

Jun-8

8

Aug-8

8

Oct-

88

De

c-8

8

Fe

b-8

9

Apr-

89

Jun-8

9

Aug-8

9

Oct-

89

De

c-8

9

Fe

b-9

0

Apr-

90

6-m

on

th %

ch

ang

e

16

18

20

22

24

26

28

30

32

Pric

e



Price(r.h.scale)

Momentum

(l.h.scale)

Peak

A

B

C

late, which gave the cycle a rightward bias. This, of course, is typicalof the second cycle in any triad. Moreover, the subsequent fall was asurprise – so much so, in fact, that it became known as the ‘bond crashof 1994’. The crash was an energy gap.

The pattern of the third sub-cycle was very much a response to thedepth of the energy gap ‘crash’. It was very muted, but it helped tofulfil the archetypal pattern, by incorporating a rally and then a declineinto the April 1996 low. Importantly, the market could not sustain aproper rally out of an oversold condition until the cycle was complete.

THE SECOND CYCLE

The second cycle within the 16-year trend cycle lasted just over fiveyears, from April 1996 to June 2001. Being the middle cycle of a trendcycle, it was particularly dynamic (see Figure 19.9). In fact, the movewas so strong that the underlying cycles were suppressed. The theo-retical timings of the sub-cycle lows, based on the completion of thewhole cycle, are indicated by the upward-pointing arrows. Quite obvi-ously, the low of wave 2 is difficult to pinpoint.

Again, however, being itself a second cycle within a triad that isscheduled to last for 16 or so years, it peaked late and unexpectedly.The subsequent fall took the market down into a 1.78-year sub-cyclelow, from where it completed the last phase of the 5.34-year cycle.


-30

-20

-10

0

10

20

30

Apr-

90

Jul-90

Oct-

90

Ja

n-9

1

Apr-

91

Jul-91

Oct-

91

Ja

n-9

2

Apr-

92

Jul-92

Oct-

92

Ja

n-9

3

Apr-

93

Jul-93

Oct-

93

Ja

n-9

4

Apr-

94

Jul-94

Oct-

94

Ja

n-9

5

Apr-

95

Jul-95

Oct-

95

Ja

n-9

6

Apr-

96

6-m

on

th %

ch

ang

e

18

23

28

33

38

43

Pric

e



Price(r.h.scale)

Momentum

(l.h.scale)

1

2

3

A

B

C

Figure 19.8 The first 5.34-year cycle within the trend cycle

CYCLE PEAKS

This simple exercise quite clearly points to the operation of a cyclethat has an average beat of 5.34 years. Not only does each beat of thecycle fit reasonably closely to the calculated rhythm but – and, in asense, more importantly – the differing patterns for each beat of thecycle adhere very closely to those predicted by the theoretical triadicmodel. The next question is, how do the peaks fit against thisconclusion?

Figure 19.10 is the same chart as that in Figure 19.4, adjusted toshow the important cycle peaks since January 1972. What is apparentis that significant peaks occur every 5.40 years. Five of the six peaksoccur at the end of sharp rallies; one cycle, however, does not. This isthe apparent peak in the middle of the March 1985 to April 1990 cycle,which occurred in March 1988. The important point, however, is thatthe actual price peak should have occurred earlier than the March1988 date because the whole cycle was the third (terminal ordetachment) cycle within the base cycle. Such cycles are usually ‘left-translated’. Nevertheless, March 1988 was the centre of the counter-trend rally that resolved into the 1989–90 bear phase.

The incidence of peaks that are 5.40 years apart obviously accordswell with troughs that are 5.34 years apart. So we can be reasonablysure that a cycle of five to five and a half years operates in the UK giltmarket.


Figure 19.9 The second 5.34-year cycle within the trend cycle

-20

-10

0

10

20

30

40

Apr-

96

Jun-9

6

Aug-9

6

Oct-

96

De

c-9

6

Fe

b-9

7

Apr-

97

Jun-9

7

Aug-9

7

Oct-

97

De

c-9

7

Fe

b-9

8

Apr-

98

Jun-9

8

Aug-9

8

Oct-

98

De

c-9

8

Fe

b-9

9

Apr-

99

Jun-9

9

Aug-9

9

Oct-

99

De

c-9

9

Fe

b-0

0

Apr-

00

Jun-0

0

Aug-0

0

Oct-

00

De

c-0

0

Fe

b-0

1

Apr-

01

Jun-0

1

6-m

on

th %

ch

ang

e

26

31

36

41

46

51

56

61

Pric

e



Price(r.h.scale)

Momentum

(l.h.scale)

1

2

3

A

B

C

FORECASTING

This gives us a starting point for forecasting future peaks and troughs.Such forecasts are dynamic (because they will be adjusted throughtime); nevertheless we can calculate initial estimates for future turningpoints. According to our model, the final peak of the disinflationarytrend cycle is likely to occur before the mid-point of its terminal cycle.That is, the third cycle within the overall disinflationary trend cycle islikely to be ‘left-translated’. We know that this high will be the fifthone along from the 1977 peak. Using the measured periodicity of thepeaks, we can therefore estimate that the mid-point of the third cycleis likely to occur 5 × 5.40 years (or 27 years) from December 1977.This is December 2004. So the price high of the last cycle in the disin-flationary trend cycle should impact before December 2004.

We can, however, be a little bit more precise. It is normal for thepeak to occur in the first one-third of this last cycle (see Chapters 15and 18). Since we are looking at a cycle that probably started in June2001 and that will likely end by October 2006, the time period for thepeak is sometime between June 2001 and March 2003.

If this last peak bears the same relationship to the whole cycle as didits April 1986 counterpart, then it is likely to impact 3 × 5.40 years (or16.2 years) from April 1986. This is June 2002. There was animportant high in September 2002.


-40

-30

-20

-10

0

10

20

30

40

50

Jun-6

9

Jun-7

0

Jun-7

1

Jun-7

2

Jun-7

3

Jun-7

4

Jun-7

5

Jun-7

6

Jun-7

7

Jun-7

8

Jun-7

9

Jun-8

0

Jun-8

1

Jun-8

2

Jun-8

3

Jun-8

4

Jun-8

5

Jun-8

6

Jun-8

7

Jun-8

8

Jun-8

9

Jun-9

0

Jun-9

1

Jun-9

2

Jun-9

3

Jun-9

4

Jun-9

5

Jun-9

6

Jun-9

7

Jun-9

8

Jun-9

9

Jun-0

0

Jun-0

1

Jun-0

2

6-m

on

th %

0

10

20

30

40

50

60

Pric

e le

vel

UK Treasury 2½% 'Irredeemable'(Monthly close and 6-month % change. Source: Datastream)

First Second Third First Second

Base Trend

Figure 19.10 Cycle peaks, January 1972 to January 1999

Of course, we cannot know for certain that a particular price high isthe high of the move. So we need to be quite clear about what mayhappen next. In particular, we need to know the critical level beyondwhich we can assume that the trend of the market has reversed. This isthe subject of the next chapter.

CONCLUSION

This chapter has provided a brief case study of cyclical behaviour in awell-known financial market. The purpose was to explain how asearch for cycles can be conducted, rather than to present a forecastfor the future as such. Nevertheless, the case study showed that, first,price patterns in the gilt market follow the pattern predicted by thetriadic cycle model of financial market behaviour; and, second, thegilt market oscillates to the rhythm of a cycle that has a central perio-dicity of 5.34 to 5.40 years. The periodicity can expand and contractthrough time – and obviously does so. However, the final arbiter of theimportant highs and lows is the pattern itself. The cycle is defined byits intrinsic pattern, not the precision of the periodicity.


270

20

Price patterns in financialmarkets

INTRODUCTION

Despite the potential accuracy of cycle analysis, it is impossible to beabsolutely sure that a turning point is occurring, even when it isactually happening. There are a number of complementary tools thatcan be used to help reduce uncertainty, and we shall look at some ofthem in more detail in Chapter 23. However, the main task is not somuch to pick the absolute high or low, but to recognize when areversal is shifting from being a correction that needs little or noresponse into being a trend that requires a change in investment posi-tions. To some extent, the difference is subjective: the boundarybetween a correction and a trend is different for a short-term traderfrom what it is for a long-term investor. For the market, however, theboundary is much more objective. The market itself ‘knows’ theprice–time frame for which it is correcting and it ‘knows’ the pointwhen it moves from one price–time frame to another (higher)price–time frame. We shall now look at this issue in more detail.

HIERARCHICAL TRENDS

As pointed out in Chapter 5, markets do not operate independently of‘fundamentals’; they operate in a limit cycle relationship with those

Price patterns in financial markets 271

fundamentals. Furthermore, such fundamentals are hierarchical innature. In Chapter 15, we briefly surveyed some of the time framesused in economics. But the scope is much broader than justeconomics: the longer the time frames that are being analysed, themore influential are factors such as socio-political trends and theclimate. Hence, we can visualize a spectrum of influences whosevibrations are very fast over short time frames and very slow overlonger ones. For example, the author and exponent of ‘the long view’Stewart Brand uses six layers of analysis, with each layer corre-sponding to a different time frame. The layers are: fashion/art;business/commerce; intellectual and physical infrastructure; gover-nance; culture; and climate.1 Undoubtedly, we could use others.However, we could say that the fashion cycle lasts about a year, thebusiness cycle lasts about three years, the infrastructure cycle takes 10years, the governance cycle may shift in rhythms of 30–50 years,culture evolves in cycles of 90–150 years, and climate oscillates inpatterns that may take 500 years and more. The parallel with some ofthe economic cycles tabulated in Chapter 15 should be obvious.

MARKETS AND FUNDAMENTAL TRENDS

For any given hierarchical level, information from a higher level islikely to impact as a shock; and the lower-level system will have toadjust to that shock. Hence, climate changes in the form of prolongeddrought or heavy rainfall can have dramatic effects on (say) a three-year business cycle. However, the emergence of a shock at the lowerlevel does not mean that the process is random. From the perspectiveof the hierarchical level from which it emerges, it is part of theprocess. It is just that, from the perspective of the lower level withlimited time horizons, the information was unexpected.

This raises some important questions about the inter-relationshipbetween market movements and changes in fundamentals. How do weknow, for example, if the market is responding to long-term deep-structure tides or to short-term surface-structure ripples?Alternatively, why should a correction that might be considered‘normal’ in the context of fundamentals suddenly turn into a blood-curdling bear that can be viewed as being ‘abnormal’ even after theevent?

ECONOMIC CYCLES, PRICE PATTERNS ANDPRICE TRENDS

One important part of the answer to these questions relates toeconomic and social cycles themselves. If we know where we are inthe bigger socio-economic cycles, then we have an idea about what toexpect from them. This is why it is so important to include an analysisof economic oscillations (at least along the lines outlined in this book)in an investment decision-making process. Another important part ofthe answer is that markets try to anticipate turning points in funda-mentals through the agency of patterned price movements. Thesepatterns are hierarchical in nature and can to some degree be inter-preted for clues about future fundamentals. The final part of theanswer is that the markets will signal when they are locking intofeedback with a higher level of fundamentals by penetrating certainboundaries.

We have already dealt with economic cycles at some length. Thenext two chapters will cover the implications of price patterns.Chapter 22 will cover the question of boundaries.

THE PATTERN OF A TREND

In Chapter 16 we concluded that there were always likely to be biasesin the three-by-three pattern of a cycle, which reflected that cycle’sposition in relation to its higher-level counterpart. One way of lookingat the consequences of these biases is to take the basic patterns and runa trend through them.

Figure 20.1 shows what happens if a rising trend is applied to a basecycle pattern. Quite clearly, wave B of the cycle can be sufficientlydistorted upwards to make a new high. As a result the cycle appears toevolve in a five-wave format: there is a base pattern, a long, rising,impulse wave, a temporary correction, and a final high. A corollary tothis is that a five-wave profile indicates that the trend is upwards.Consequently, the emergence of five waves in a base cycle can be usedto confirm that the trend has changed. The signal would be givenwhen the new high occurs within wave B. Following such a signal, theinference is that setbacks are buying opportunities.


THE FIVE–THREE WAVE PATTERN INCOMPLEX STRUCTURES

Very importantly, however, the five-wave profile is not just confinedto base cycles. It also emerges in the context of more complex struc-tures. Figure 20.2 shows what happens if a rising trend is applied tothe more complex cycle pattern that incorporates sub-cycle biases.Ignoring the very low-level fluctuations, a very important patternemerges. First, the base cycle develops five waves (i–ii–iii–iv–v) anda short correction (from 1 to 2). Second the long, impulsive trendcycle embodies five waves (i–ii–iii–iv–v) and a limited correction(from 3 to 4). Third, the terminal cycle evolves as a sharp rally to anew high (at 5) followed by a three-wave correction (A–B–C). Thewhole pattern therefore seems to contain three rising impulse waves(1, 3 and 5, of which 3 is the longest), two corrections (2 and 4) and athree-wave fall (A–B–C). In other words, a bull market plus a naturalcorrection presents itself as a five–three pattern.


Figure 20.1 Base cycle plus trend

THE ELLIOTT WAVE PRINCIPLE

The idea that markets move in a five–three pattern was the perceptiveinsight of the great stock market theorist Ralph Nelson Elliott.2

Writing in the 1930s, Elliott concluded that the five–three profile wasessentially a law of nature. He deduced that the associated patternsobeyed certain clearly defined rules that could be used in trading. Heobserved, however, that not all patterns actually matched the basicfive–three profile, and he accordingly went to great lengths to showthat any aberrations were caused by distortions to the underlyingfive–three pattern. In particular, he argued that the fifth wave of animpulse movement could ‘fail’, so that the whole movement essen-tially looked like a three-wave movement; and he argued that anapparent fifth wave that consisted only of three sub-waves was part ofa correction.

Elliott’s ‘wave principle’ provides a wonderfully thoroughcoverage of price patterns in financial markets. Indeed, there is noknown pattern that falls outside of his framework. The truth, however,


Figure 20.2 Complex cycle plus trend

is that the five–three pattern is not a fundamental law of nature. It is aderivative of a much more fundamental three–three cycle pattern,which is a law of nature.

Nevertheless, Elliott’s findings are a valid and practical philo-sophical approach to markets, and his ‘wave principle’ invokes rulesthat are valid interpretations of the implications of the three–threeprice pulse. Traders who use the Elliott wave principle will generallyobtain exceptionally good results. For that reason, we shall analyseElliott’s conclusions in a little more detail in the next chapter. In themeantime, however, we need to assess the implications of thethree–three price pulse in a little more detail.

INVESTMENT GUIDELINES

It should be apparent from Figure 20.2 that there are 18 waves in acomplete move – namely six waves within three sub-cycles. However,distortions to the natural cycle allow variations both in the pattern of arally and in the pattern of a fall. A rally can consist either of threewaves (when the cycle doesn’t have an uptrend) or of five waves(when it does). Meanwhile, a correction can consist either of one wave(if the cycle has an uptrend) or of three waves (if it doesn’t). Thisknowledge implies four very important guidelines that can be used forinvestment and trading. The first is the buy signal after a low. Thesecond is the warning that a bear phase has started. The third relates towhen a correction can be expected. And the fourth involves themessage that can be derived from the strength of contra-trend rallieswithin corrections.

THE BUY SIGNAL AFTER A LOW

We have already mentioned the importance of a five-wave rally after alow. This would be an indication that the trend has turned upwards. Ifthis is true, then the correction immediately following the five waves,which completes the first (base or detachment) cycle, is going to bevery limited. Consequently, the next new high is going to provide aclear buy signal because the implication is that the market has movedinto its second (trend or transition) cycle. This latter has the potentialto be the most dynamic (see Figure 20.3).


WARNING OF A BEAR PHASE

Because the basic pattern of the price pulse is three-up/three-down,it follows that patterns that emerge in rising markets are also likelyto materialize – although in reverse – in falling markets. Hence, ifthe presence of five waves in an upward direction signals the possi-bility of a bull trend, the emergence of five waves in a downwarddirection will signal the possibility of a bearish one. This is demon-strated in Figure 20.4. Here, a downtrend is applied to the basic triadof cycles, starting from the peak of the first cycle in the triad. Inother words, it is assumed that the market comes to a natural cyclicalpeak and then starts correcting downwards, but that the correction isactually for a higher-level cycle. The result is that what should havebeen the base phase in the second cycle is so biased downwards thatit makes a new low.


Figure 20.3 The buy signal

THE TERMINAL CYCLE

This is basically what is likely to happen in a terminal cycle. Themarket (or economy) peaks during the first cycle in a triad and thenmakes a low in the early part of the second cycle. In other words, thenatural cycle low at the end of the first cycle provides support but nota final turning point. We provided evidence of this phenomenon inChapter 18. The difficulty, however, is that the five-wave pattern maynot necessarily impact quickly enough to provide a strategic sellsignal. Obviously, a trading signal is given as wave 5 makes a newlow; but it stands to be quickly reversed as wave 3 of the second cyclesnaps in. The evidence is that there is then normally a long three-waverally prior to a final – and often very difficult – down-wave right at theend of the third cycle (see Figure 20.5). Good examples of thisphenomenon were shown in Figure 18.5 in Chapter 18.


Figure 20.4 The five-wave downtrend

THE LOCATION OF CORRECTIONS

The second important guideline is that a contra-trend correction willoccur after a five-wave rally. In the simplest case, the correction willjust consist of a simple down-wave. In the more complex case, thecorrection will consist of three clear waves – ie, two falling waves,interspersed with a rising wave.

The simple down-waves are likely to occur when a market istrending strongly. Moreover, because it derives from wave C of thecycle, it is likely to be larger in price terms than either of the twopreceding corrections (which would have been wave 2 and wave A). Itis relevant, however, that it will probably encapsulate three lower-order waves (see Figure 20.6). It will therefore take on a three-wave‘look’ (ie, a–b–c).

The more complex down-waves are likely to occur within thecontext of big, longer-term cycles that are losing their impetus (seeFigure 20.7). Since it consists of three distinct waves (A–B–C), such acorrection is likely to be greater in both price and time than either ofits two predecessors. Further, once account is taken of the unfolding oflower-order waves, it is easy to see that each wave within thecorrection will itself consist of three waves.


Figure 20.5 The pattern of the terminal cycle


Figure 20.6 Simple corrections

Figure 20.7 Complex corrections

THE STRENGTH OF CONTRA-TREND RALLIES

As presented in Figures 20.2 and 20.7, the final three-wave correction(A–B–C) is assumed to take on a ‘zigzag’ appearance. This will allowquite a deep correction. Sometimes, however, the contra-trend rally isquite strong, taking it up very close to the wave 5 high, so that it has a‘flat’ appearance (see Figure 20.8). Strong contra-trend waves areprima-facie evidence that the higher-order uptrend is still intact. Forexample, it may be part of a higher-order base (or detachment) cycle.Hence ‘double tops’ are likely to be penetrated.

NON-CONFIRMATION REVISITED

It is worth at this point commenting again on the ‘non-confirmation’of new highs or new lows. We referred to this concept in Chapter 13.Basically, non-confirmation means that a new price high or a newprice low is not fully reflected in one or more of the auxiliary indicesthat measure the energy of a trend.


Figure 20.8 Flat correction

The foregoing analysis indicates that wave B in a rising market, andwave 2 in a falling market, may be forced into new territory for thetrend. In principle, however, both are contra-trend waves and are notintrinsically energetic. Hence, a properly defined energy indicatorshould pick up on this flagging energy. One of the most relevant, andeasiest to use, is a simple momentum indicator. Shown in Figure 20.9is the weekly close of the US dollar in terms of Japanese yen for theperiod September 1999 to November 2002. It reveals a five-wave rallysince the November 1999 low. Also shown on the chart is the 52-weekpercentage change in the weekly closes. For most of the rally, themomentum indicator moves to a peak of 3 to 4 per cent and then turnsdown as a contra-trend correction develops. By the time the dollarreached the final high in the five-wave move, however, the momentumindicator was dropping sharply. This was a genuine non-confirmation.

Of course, a non-confirmation using a simple rate of change indi-cator needs to be backed up by other tools. We shall have a very brieflook at some of these extra tools in Chapter 23.

In the meantime, however, there is obviously an important inferencethat we can draw from this analysis. This is that a genuine non-confirmation – that will be associated with a significant correction –will only arise in a ‘fifth wave’.


80

90

100

110

120

130

140

Se

p-9

9

No

v-9

9

Ja

n-0

0

Ma

r-0

0

May-0

0

Jul-00

Se

p-0

0

No

v-0

0

Ja

n-0

1

Ma

r-0

1

May-0

1

Jul-01

Se

p-0

1

No

v-0

1

Ja

n-0

2

Ma

r-0

2

May-0

2

Jul-02

Se

p-0

2

Exch

ang

e ra

te

-5

0

5

10

15

20

Per

cen

t p

.a.

Dollar in terms of Yen(Weekly hi/lo/close)

Source: Bloomberg

1

2

5

4

3

Exchange rate(l.h.scale)

Momentum(r.h.scale)

No

v-0

2

Figure 20.9 Non-confirmation in a fifth wave

ANTICIPATING TURNING POINTS

There is one further piece that needs to be added to the jigsaw puzzle.In Chapter 11, it was argued that turning points could be anticipatedwith the help of the golden ratio. Specifically, it was shown thatimpulsive movements could be related to the correction that precededthem by the ratio 2.618:1, which is a derivative of the golden ratio.

The ratio 2.618:1 is undoubtedly the most important ratio in calcu-lating price objectives. In particular, it tends to apply to wave 3, whichis often the longest wave, of a five-wave movement. However, it is notthe only relevant ratio. The main point is that waves of any degreetend to be extended or damped by the influence of higher-order trends.However, even here, it seems to be a feature of nature that golden ratioinfluences still operate because price movements remain mathemati-cally related to one another. One of two things may happen. First, ahigher-order trend may counteract the impulse wave, so that theappropriate ratio for calculating the subsequent objective is 1.618:1.This often happens either in the last wave of a five-wave impulsemovement or in the last wave of a three-wave counter-trend move.Second, a higher-level trend may complement the impulse wave, sothat the appropriate ratio is 4.236:1. This latter may be found in partic-ularly dynamic impulse waves.

These ratios are all related to each other by the golden ratio. Theycome from the (infinite) number series

… 0.236, 0.382, 0.618, 1, 1.618, 2.618, 4.236…

where each number in the series is 161.8 per cent higher than thenumber that precedes it.

THE GOLDEN RATIO FORMULAE

We have already dealt with the 2.618 formula in some detail inChapter 11. We can now place it in a more general context. Figure20.10 shows a schematic diagram for a bull move, whereby the depthof a correction is multiplied by 1.618, 2.618 and 4.236 and theresulting outcomes are added to the level at which the correctionended. Reversals can then be expected after a wave has travelled161.8 per cent, 261.8 per cent or 423.6 per cent of the depth of thecorrection. Corresponding calculations can also be made for a fallingmarket. Here, the extent of a contra-trend rally is multiplied by each of


the numbers 1.618, 2.618 and 4.236. The resulting figures arededucted from the level that marks the top of the counter-trend rally.

SOME EXAMPLES

Shown in Figure 20.11 is the behaviour of the US dollar in terms of theJapanese yen between June 1998 and June 2002. We shall just pick outsome of the main movements. Hence, the last up-wave of the toppattern between June and August 1998 was ¥13.97. Multiplying thisby 2.618 and deducting the result from the high of ¥147.66 generateda target of ¥111.08. The actual temporary low in October 1998 was¥111.58.

Between early June and early July 1999, the market traced out amini top pattern by retracing part of the initial fall from the mid-May1999 high. The rally was only ¥5.00. However, multiplying that by4.236, and then deducting that from the top of the re-test at ¥120.90,


Figure 20.10 Golden ratio calculations

produced a downside target of ¥101.75. The market reached a finallow at ¥101.25.

The US dollar then formed a base between late November 1999 andlate March 2000. The re-test of the low, in March, was ¥9.66.Multiplying this by 2.618 and adding the result to the re-test low of¥102.07 produced an objective of ¥127.35. The market rallied to¥126.85 in March 2001.

Between March and September 2001, the dollar fell by ¥11.01.Multiplying this by 1.618 and adding the result to the end of thecorrection produced a target of ¥133.64. The market reached ¥135.14in late January 2002.

These examples are taken from just one market over a relativelylimited period of time. It was easy to find that each of the mainimpulse movements bore a golden ratio relationship to the contra-trend move from which it emerged. It was not necessarily possibleprior to the event to know which multiple would be applicable to themove, and there is quite obviously always some room for manoeuvrein the precision of the calculations.

Nevertheless, it would be possible to produce literally thousands ofother examples of this extraordinary inter-relationship betweenpatterns and ratios. The strong evidence is that the underlyingprocesses in financial markets are not random.


95

105

115

125

135

145

Ju

n-9

8

Au

g-9

8

Oct-

98

De

c-9

8

Fe

b-9

9

Ap

r-9

9

Ju

n-9

9

Au

g-9

9

Oct-

99

De

c-9

9

Fe

b-0

0

Ap

r-0

0

Ju

n-0

0

Au

g-0

0

Oct-

00

De

c-0

0

Fe

b-0

1

Ap

r-0

1

Ju

n-0

1

Au

g-0

1

Oct-

01

De

c-0

1

Fe

b-0

2

Ap

r-0

2


Source: Bloomberg

261.8% = 127.35

261.8% = 111.08

423.6% = 101.75

161.8% = 133.64

Ju

n-0

2

Figure 20.11 Golden ratio calculations for the dollar–yen

CONCLUSION

In this chapter, we have looked at the way that price patterns canprovide a clue to the nature of the underlying trend. We have foundthat the emergence of a five-wave pattern is a strong indication that atrend is in force. We have also found that the price span of a five-wavetrend may to some extent be predictable using the golden ratio and thata five-wave pattern is followed by a correction. Quite obviously thehigher the level at which the five-wave pattern occurs, the deeper willbe the associated correction. We therefore need to look more closely atthe nesting of price patterns; and we shall do so by surveying RalphNelson Elliott’s ‘wave principle’.

NOTES

1. Brand, Stewart (1999) The Clock of the Long Now: Time and responsibility,Weidenfeld & Nicolson, London.

2. Elliott, Ralph Nelson (1946) Nature’s Law: The secret of the universe, Elliott,New York, reprinted in Prechter, Robert (ed.) (1980) The Major Works of R NElliott, New Classics Library, New York.


286

21

The Elliott wave principle

ELLIOTT’S DISCOVERY

The Elliott wave principle is a widely used, but little understood,philosophical approach to stock markets. During a long period ofconvalescence in the early 1930s, Elliott occupied himself by under-taking a detailed analysis of stock market movements. In the process,he discovered a unique formula that ‘defined’ stock market move-ments. This formula was based on the premise that the stock marketaverages rise in five ‘waves’ and fall in three ‘waves’, and it subse-quently enabled Elliott to predict price movements with a degree ofaccuracy that astounded contemporary commentators. In 1938, hepublished his findings in his book entitled The Wave Principle.1

The Elliott wave principle has been popularized by Robert Prechterin the United States in recent years, and its enviable record of successhas created a large following. Nevertheless, it is fair to say that thewave principle suffers from two specific problems:

� it is an extraordinarily complicated system to apply;

� no one (including, it has to be said, Elliott himself) has actuallybeen able to explain why the central formula works.

The Elliott wave principle 287

THE PRICE PULSE AS THE BASIS OF THEWAVE PRINCIPLE

As we have demonstrated, the key to understanding Elliott’s formulais the price pulse. Once this is recognized, the whole of Elliott’sanalysis becomes an acceptable method of analysing, and forecasting,stock market prices. Quite simply, price pulses from different levels ofthe hierarchy combine with one another, thereby creating the differentpatterns that Elliott found conformed to certain simple rules. Let ustherefore survey Elliott’s findings in more detail.

THE BASIC WAVE PATTERN

Elliott’s primary assertion was that all bull markets consist of fivewaves and that all bear markets consist of three waves. These areshown in Figure 21.1. Each bull phase therefore consists of threeimpulse waves (1, 3 and 5 in Figure 21.1), interspersed with twocorrective waves (2 and 4 in Figure 21.1). The bear phase consists oftwo impulse waves (A and C in Figure 21.1) interspersed with onecontra-trend corrective wave (B in Figure 21.1).

Figure 21.1 The Elliott wave

CORRECTIONS

In this simple formulation, waves A, B and C of the bear phase can bebroken down into easily identifiable patterns. Waves A and C areimpulse waves. Elliott considered that both would therefore consistof five waves. Wave B, on the other hand, is a corrective wave (thatis, it corrects during a correction). It therefore consists of threewaves, and is an inverted version of the whole A–B–C correctionitself (see Figure 21.2).

A UNIVERSAL PHENOMENON

An important feature of the Elliott wave principle is that it applies toall degrees of movement. Hence, waves 1, 3 and 5 of any five-waveimpulse movement each consists of five waves itself, while waves 2and 4 of that movement each consist of three waves. Hence, eachcomplete five–three bull–bear cycle becomes part of the cycle of thenext higher degree. For example, the bull–bear formation shown inFigure 21.1 could constitute either waves (1) and (2), or waves (3) and(4), of a higher-order cycle2 (see Figure 21.3). In this sense, Elliott’sprinciple embraces a theory of growth: ultimately, bear market pricefalls will be completely reversed, and the previous bull market pricepeak will be exceeded.


Figure 21.2 Basic Elliott corrective wave

THE WAVE PRINCIPLE AS A NATURALPHENOMENON

Elliott therefore made the important observation that stock marketbehaviour was a natural phenomenon that contained its own intrinsicpatterns. This, of course, had fundamental implications for the processof forecasting stock market behaviour, because, in principle, thepatterns of the lower levels should regenerate themselves at a higherlevel.

DERIVED RULES: TREND INDICATIONS

Elliott, in fact, established certain rules by which it would be possibleto both establish the direction of the main trend in the market anddetermine when a market reversal, or turning point, was in the processof occurring. First, he noted that the emergence of a five-waveimpulse pattern, either upwards or downwards, would provide astrong indication of the direction of the longer-term trend. Hence, arising five-wave pattern after a sharp fall in the market would besuggestive of further rises; while a falling five-wave pattern after asharp rise would suggest that further falls would take place.


Figure 21.3 The complete Elliott wave

DERIVED RULES: IMPULSE WAVES

Second, within each five-wave movement, three basic rules wouldapply:

� wave 4 will not penetrate below the peak of wave 1;3

� wave 3 is often the longest, but is never the shortest, of the fiveimpulse waves that constitute the whole movement;

� two of the three impulse waves will be equal in length.

DERIVED RULES: CORRECTIONS

Third, on the question of corrections, Elliott noted three basicguidelines:

� no A–B–C formation will ever fully retrace the preceding five-wave formation to the same degree;

� each correction will be at least as large in price, and as long in time,as all lower-degree corrections that preceded it;

� each correction tends to return to the price range spanned by acorrective wave of one degree lower – that is, either to wave 2 or 4.4

COMPLICATIONS WITHIN THE SYSTEM

Elliott’s basic five–three formulation, together with its associatedguidelines, provided a unique view of the way that stock marketsoperate. Unfortunately, however, the formula is not totally compre-hensive. Elliott himself found that although the basic five–threepattern applied to a large number of situations, there were, never-theless, certain variations that had to be taken into account. Whenthese variations were included, then indeed the analysis wascomplete: there is no known stock market price pattern that fallsoutside Elliott’s framework.


FIFTH-WAVE VARIATIONS:FAILURES AND EXTENSIONS

The first set of variations that Elliott dealt with were those that applyto the fifth wave of an impulse movement. The basic formulationsuggests that the fifth wave will travel beyond the end of the thirdwave of the same degree. However, in practice the fifth wave caneither fall short of the end of the fifth wave (a failure) or extend itselfin an additional dynamic five-wave movement (an extension). Theseare shown in Figures 21.4 and 21.5 respectively.

BEHAVIOUR FOLLOWINGFAILURE OR EXTENSION

The occurrence of either a failure or an extension in the fifth wave of amovement gives a very clear indication of the subsequent behaviour


Figure 21.4 Fifth-wave failures

Figure 21.5 Fifth-wave extensions

of the market. A fifth-wave failure is obviously the sign of greatweakness at the end of a bull run or of great strength at the end of abear phase. Corrections following a fifth-wave bull market failure willtherefore be very deep in price terms; rallies following a bear marketfailure will be very sharp. These are shown in Figure 21.6.

On the other hand, extensions signal a very dynamic impulsemovement.5 Consequently, in a bull market, prices will move to newhigh ground even after the apparent peak of the fifth wave; in a bearmarket, prices will drop to new lows after the apparent end of the fifthwave. These are shown in Figure 21.7. Elliott called this price actionthe ‘double retracement principle’ as prices would twice cover theground traced out by the extension itself. The first retracement wouldreturn to the beginning of the extension (shown as ×2 in Figure 21.7),and the second retracement would move the market back past the endof wave 5.6


(a) End of bull phase ( b) End of bear phase

Time

B

C

Figure 21.6 Movement following fifth-wave failure

Figure 21.7 Movement following fifth-wave extension

FIFTH-WAVE VARIATIONS: DIAGONALTRIANGLES

In addition, Elliott found that sometimes the fifth wave consists of adiagonally sloping triangular formation where:

� the formation consists of five waves;

� each wave is a three-wave movement;

� the fourth wave may penetrate below the peak of the first wave.7

This pattern is relatively uncommon, but is always followed by adramatic move in prices in the opposite direction to the slope of thetriangle. Two examples are shown in Figure 21.8.

VARIATIONS IN CORRECTIONS:THE THREE-PHASE A-WAVE

The second set of variations analysed by Elliott relates to the form ofthe corrective patterns. The basic pattern, which was reflected inFigure 21.2, is called a ‘zigzag’ – that is, it consists of two decliningwaves interspersed with a contra-trend rally. As explained earlier, boththe A- and the C-waves of a zigzag movement consist of five decliningwaves, and the B-wave consists of three waves. The zigzag correctionis therefore a five–three–five formation. The variations from the basiczigzag pattern relate to the fact that the A-wave often consists only ofthree waves, thereby yielding a three–three–five formation.8 The B-wave following such a ‘truncated’A-wave tends to be very strong –


Figure 21.8 Fifth-wave diagonal triangles

so much so, in fact, that it travels at least to the point at which the A-wave started. Elliott actually found three possible three–three–fiveformations. These he called the ‘regular flat correction’, the ‘irregularflat correction’ and the ‘running correction’ respectively.

THE FLAT CORRECTION

With a regular flat correction (see Figure 21.9), the B-wave takesprices back to the level they were at at the start of the A-wave, and thesubsequent C-wave returns prices to the level they were at at the endof the A-wave. In other words, prices fluctuate within a trading rangebefore resuming their trend.

With an irregular flat, however, there are two distinct alternatives.First, the B-wave may take prices back above the level at which theywere at the start of the A-wave. Second, the C-wave may either fallsignificantly short of the bottom of the A-wave or it may significantlyexceed the bottom of the A-wave. Two of the possible combinationsare shown in Figure 21.10.

In extreme cases, the relatively rare running correction occurs.Here, the B-wave substantially exceeds the start of wave A, and thesubsequent C-wave also leaves prices above the start of wave A (seeFigure 21.11).


Figure 21.9 Regular flat correction

Elliott’s basic rule concerning both flat and running corrections wasthat they are all indicative of the fundamental trend in the market.Specifically, the strength of the B-wave is an important indication thatmarket prices will rally very sharply following the end of a correctionand that prices will probably exceed the peak of that B-wave.

COMPLEX CORRECTIONS

Obviously, therefore, corrections are potentially complicated move-ments. Their essential purpose is to counteract the excesses of theprevious rising impulse wave of the same degree and to prepare theconditions for the next rising impulse wave. Sometimes, however, asimple zigzag or flat correction is insufficient for this task.


Figure 21.10 Irregular flat corrections

Figure 21.11 Running correction

Consequently, the market extends the correction by combining two ormore corrective patterns. In principle, any number of such patternsmay be strung together, and the only restriction on the resultingcomplicated formation is that each A–B–C pattern is separated fromthe next one by another A–B–C pattern. This latter intervening patternis usually called an X-wave. Two examples of complex correctionsare shown in Figure 21.12.

TRIANGLES

One of the complex corrections that Elliott considered to be particu-larly important was a triangular formation. Such a formation essen-tially consists of five A–B–C waves or, more precisely, three A–B–Cwaves interspersed with two X-waves – and the extent of travel ofeach successive A–B–C movement is contained by boundary linesthat are either contracting or expanding.9 Relevant examples areshown in Figure 21.13.

THE IMPLICATIONS OF A TRIANGLE

The importance of the triangle lies in the fact that it gives a uniqueindication of the subsequent behaviour of market prices:

� market prices exit the triangle in the same direction as wave 2 ofthe triangle;


zag

Figure 21.12 Complex corrections

� the resulting impulse wave is the last such wave in the direction ofthe main trend;10

� in the case of contracting triangles, prices tend to move by anamount equal to the widest part of the triangle and to change trendin line with the apex of the triangle.11

These rules are reflected in the example shown in Figure 21.14.


Figure 21.13 Triangular corrections

Figure 21.14 Price movement after a triangle

INVERTED CORRECTIONS

Although all these comments relate to corrections within the contextof a rising market, they also apply to contra-trend movements within afalling market. Hence, they apply to waves 2 and 4 of a falling impulsewave and to the B-wave of an ordinary correction. It is usual to applythe term ‘inverted’ to the relevant formation. Examples of an invertedzigzag, an inverted regular flat, an inverted irregular flat and aninverted symmetrical triangle are shown in Figure 21.15.

THE ‘RULE OF ALTERNATION’

In his book Nature’s Law, Elliott argued that an essential feature of thewave principle was the ‘rule of alternation’. At its simplest level, alter-nation can be taken to mean that bear markets alternate with bullmarkets. More importantly, however, Elliott argued that successive


Inverted zigzag

Figure 21.15 Inverted corrections

corrective patterns will be different from each other. Hence, within afive-wave impulse movement, wave 2 will differ from wave 4 both intype and complexity. Within a complicated correction, adjacent three-wave patterns will also differ from one another, both in type andcomplexity. Accordingly, for example, a zigzag might be followed bya flat, and a simple correction will be followed by a complicatedcorrection consisting of two or more three-wave patterns. However,Frost and Prechter12 suggest that the tendency for corrections toalternate should be regarded as a probability rather than an inviolablerule. There are actually numerous examples where alternation doesnot occur.

THE PROBLEMS WITH THEELLIOTT WAVE PRINCIPLE

Despite the fact that the Elliott wave principle provides a completecatalogue of price patterns, many analysts find it very unsatisfactory.There are four interrelated reasons for this:

� There has previously been no satisfactory explanation of thefive–three formation. Elliott himself considered that the formationwas a ‘law of nature’13 and left it at that. Unfortunately – and apartfrom the fact that the numbers derive from the Fibonacci sequence– the five–three pattern does not obviously appear as a regularphenomenon anywhere else in nature. It is difficult to believe thatsomething that is supposedly so fundamental is both totally inex-plicable and without parallel.

� The integrity of the central theorem of a five–three pattern for eachbull–bear cycle is only maintained by the use of the concepts offailures and irregular corrections. This suggests the absence of aunifying cause and effect.

� The complexity of the Elliott wave (particularly in so far as exten-sions and complicated corrections are concerned) means that it isoften impossible to establish a unique future out-turn for marketprices.

� It is possible to isolate alternatives (and therefore probabilities) inthe knowledge that any actual out-turn will be a valid expressionof the Elliott principle, but precise forecasts are difficult to makeprior to the event.


Furthermore, even those analysts who accept the basic validity of theprinciple, and therefore make substantial use of it, have noted someadditional problems:

� The principle is more difficult to apply to some markets than toothers – for example, it is difficult to apply to commodity marketsbecause wave 4 of a rising impulse wave often penetrates belowthe peak of wave 1 of the same degree.

� It is also difficult to apply to foreign exchange markets where afive–three pattern for one currency in a particular cross-rate consti-tutes a three–five pattern for the other currency in that cross-rate.14

� The wave principle suggests that the stock market will always rise:each upward wave is always part of a larger upward wave, and allcorrections will terminate above the start of wave 1 of thepreceding bull run, but this conclusion implies that dynamicsystems grow quantitatively, but do not change qualitatively,which is at odds with the basic principles of natural development.

All these difficulties disappear once the implications of the price pulseare understood. Specifically, all movements unfold in three wavesunless there is a qualitative (that is, second-order) change. In this lattercase, the movement takes on a five-wave profile. Elliott’s five–threewave principle is therefore actually a product of an evolving economy.

NOTES

1. Elliott, Ralph Nelson (1938) The Wave Principle, Elliott, New York, reprintedin Prechter, Robert R (ed.) (1980) The Major Works of R N Elliott, NewClassics Library, New York.

2. In differentiating between movements of differing degrees, it is conventionalto use different notations. For example, the lowest-level waves are usuallynumbered using the lower-case Roman numerals i, ii, iii, iv, v, followed by a, b,c. The next level is (i), (ii), (iii), (iv), (v), followed by (a), (b), (c). Then comes1, 2, 3, 4, 5, followed by A, B, C. Then (1), (2), (3), (4), (5), followed by (A),(B), (C). And so forth.

3. Except in cases of relatively rare triangular formations and commoditymarkets.

4. Indeed, Elliott found that, very often, a correction would return to the lower-degree corrective wave that bore the same relationship to the larger movement.Hence, a higher-degree wave (2) would return to the lower-degree wave 2,while a higher-degree wave (4) would return to the lower-degree wave 4.

5. Elliott argued that fifth-wave extensions were likely to occur if waves 1 and 2were short and equal in length.


6. According to the price pulse formulation, an extension during (for example) arising market emerges during wave 3. The double retracement is then causedby the subsequent wave A and wave B, but where wave B itself moves to a newhigh. The whole pattern assumes a sharply rising trend. The converse is so fora bear market.

7. In stock markets, this is the only exception to the rule that wave 4 correctionscannot penetrate below the end of wave 1 of the same degree. The evidencesuggests that the rule is also sometimes violated in commodity markets.

8. This implies that all movements end on a five-wave count. However, Elliotthimself was not entirely convinced by this, and he dabbled with the concept ofan ‘A–B base’ that was a three–three construct. See Elliott, Ralph Nelson(1946) Nature’s Law: The secret of the universe, Elliott, New York, reprinted inPrechter, Robert R (ed.) (1980) The Major Works of R N Elliott, New ClassicsLibrary, New York. The A–B base is a perfect reflection of the wave 1 andwave 2 base of the price pulse.

9. The boundary lines of the triangle need not converge nor diverge in a symmet-rical fashion. In an ascending triangle, the top boundary line is horizontal,while in a descending triangle, the bottom line is horizontal.

10. That is, in a bull run, triangles can only occur in wave 4.11. Frost, Alfred J, and Prechter, Robert R (1978) Elliott Wave Principle, New

Classics Library, New York.12. Ibid.13. Elliott, Ralph Nelson (1946) Nature’s Law: The secret of the universe, Elliott,

New York, reprinted in Prechter, Robert R (ed.) (1980) The Major Works of R NElliott, New Classics Library, New York.

14. This problem does not necessarily exist where the US dollar is being used asthe base currency. This is because the dollar is an international medium ofexchange and store of value.


302

22

Information shocks andcorrections

INTRODUCTION

The pattern of prices can help market participants to gauge where theyare in the context of the bigger trends. However, the art is always to beable to stay with the trend long enough to make a serious profit andthen to close positions without giving too much of that profit back.This, again, means that it is important to be able to distinguishbetween a technical correction that need not involve the wholesalereversal of investment positions and a fundamental trend reversal thatdoes require such a change. The great secret is not only that the market‘knows’ the location of these important boundaries, but also that itconveys this information to those who are willing to look for it. Weneed, therefore, to look more closely at information shocks and themarket’s response to those shocks.

INFORMATION SHOCKS

In Chapter 9, we pinpointed two types of information shock that arerelevant to the discussion. These shocks occur:

� when the market recognizes that it has changed direction; and

� when the market is accepting an extension of its trend.

Information shocks and corrections 303

The former is likely to occur at, or just after, the beginning of wave 2in a bull phase or the beginning of wave C in a bear phase. The latter islikely to occur towards the end of wave 3 in a bull phase or towardsthe end of wave C in a bear phase. These shock points are likely to beaccompanied by a surge in the market’s energy indicators: volumeswill increase, price gaps will appear, daily price spreads betweenhighs and lows will widen and short-term momentum will rise.

INFORMATION SHOCKS AND AFIVE-WAVE TREND

Shock points indicate moments when the market accepts that a changehas occurred. They subsequently become very important points ofreference, which can be used to calculate entry and exit levels. Weshall look at these calculations shortly. Before doing so, however, weneed to explore the role of information shocks in the unfolding of thefive–three pattern of price trends. First, in any impulsive five-wavepattern, there will be three shock points, which either initiate orrestore the trend. The first shock point will signify the change in trendand will be associated with an energy gap, and the other two shockpoints will signify the end of contra-trend corrections and/or thebeginning of an impulse wave.

Second, in any five-wave trend there is likely to be at least one pro-trend shock point, which signifies second-order change. This willlikely emerge in the vicinity of what might otherwise have been anenergy gap. In other words, to create a five-wave pattern, a pro-trendshock may be needed to over-ride an energy gap.

The relevant shock points in the context of a bull trend are shown inFigure 22.1. Those relating to a bear phase are shown in Figure 22.2.

SHOCKS AND CYCLES

In Figures 22.1 and 22.2, the timings of contra-trend lows that followan information shock coincide with the natural rhythm of underlyingcycles. Figure 22.3 shows such a situation in the context of thesterling–dollar exchange rate. Sterling made an important low againstthe US dollar in June 2001. The rate rallied until October 2001, beforere-testing the low in February 2002. Using the time elapse betweenJune 2001 and February 2002 (ie, roughly 32 weeks) as indicating the


Figure 22.1 Information shocks in a bull trend

Figure 22.2 Information shocks in a bear trend

influence of a cycle, the high of cycle wave 3 was due in May 2003.However, just after getting there, the market generated an importantbuy signal. So, instead of entering a wave B contraction, the marketjumped the energy gap and surged into a bull trend. Nevertheless, theunderlying cycle seemed still to operate because lows were reached inAugust 2002 and April 2003. In other words, the shock that allowedsterling to jump the energy gap did not distort the underlying cycles.

On the other hand, a shock may derive from a sufficiently high levelof the cycle hierarchy to upset the rhythms of the lower levels. Figure22.4 shows the impact of a shock, in the context of a (roughly) 22-week cycle, on the Dow Jones Industrial Average. The Dow made amajor low in September 2002, rallied into the turn of the year, andthen made a re-test low in February 2003. The high of the cycle wasdue in May 2003 but that high took the market out of the trading rangeof the base pattern. So, the market jumped the energy gap and rose intoa bull trend. This trend was sufficiently strong to damp down (but noteliminate) the underlying cycle for a long time – probably, in fact,until at least October 2004.

One way of thinking about this phenomenon is to consider whathappens to the human heartbeat when the body moves from walkingto running: the result is a much faster heartbeat. That is, the cycleadjusts to the information shock, rather than the other way around. Infinancial markets, pro-trend shocks that come from one level of thehierarchy may not have much of an obvious distorting effect on thecycles on the next lower level, other than through a slight expansion or


Figure 22.3 Information shock and underlying cycles

1.4

1.45

1.5

1.55

1.6

1.65

1.7

Dec

-00

Jan-

01

Feb-

01M

ar-0

1

Apr

-01

May

-01

Jun-

01

Jul-0

1

Aug

-01

Sep

-01

Oct

-01

Nov

-01

Dec

-01

Jan-

02

Feb-

02M

ar-0

2

Apr

-02

May

-02

Jun-

02

Jul-0

2

Aug

-02

Sep

-02

Oct

-02

Nov

-02

Dec

-02

Jan-

03

Feb-

03M

ar-0

3

Apr

-03

May

-03

Jun-

03

Jul-0

3

Aug

-03

Sep

-03

Exch

ange

rate

Sterling–Dollar(Weekly, hi/lo/close)


1 3 B

A2 C

Shock

contraction in timings. However, such shocks may have quite an effecton cycles that are two or more levels below. Hence, a pro-trend shockderiving from a 12-year cycle may not necessarily upset the basicrhythm of a 4-year cycle, but it could have quite a profound impact onunderlying 1.33-year cycles and below. The beat of the 1.33-yearcycle may be so extended that it hardly registers or is totally elimi-nated. Indeed, a large-enough information shock will need to beabsorbed independently of lower-level cyclical fluctuations, and thismay even give the impression of randomness. As far as the analyst isconcerned, it is much more important to watch the pattern after a priceshock than it is to rely on the precision of cycle periodicities.

INFORMATION SHOCKS AND BOUNDARIES

A five-wave upswing is the signature of second-order change, or ofevolution, while a five-wave downswing is the signature of a break inthe uptrend and/or of involution. Neither should be ignored. Despitethe compelling nature of the five-wave pattern, however, there willstill be some uncertainty about when the market is actually beginningto trend. Technically, a trend only really starts when the market beginsto lock into a higher level of changing fundamentals. Prior to thispoint, the market is only correcting within the context of unchangedfundamentals.


7500

8000

8500

9000

9500

10000

10500

11000

11500

Jun-

02

Jul-0

2

Aug

-02

Sep

-02

Oct

-02

Nov

-02

Dec

-02

Jan-

03

Feb-

03

Mar

-03

Apr

-03

May

-03

Jun-

03

Jul-0

3

Aug

-03

Sep

-03

Oct

-03

Nov

-03

Dec

-03

Jan-

04

Feb-

04

Mar

-04

Apr

-04

May

-04

Jun-

04

Jul-0

4

Aug

-04

Sep

-04

Oct

-04

Nov

-04

Dec

-04

Jan-

05

Inde

x le

vel

Dow Jones Industrial Average(Weekly, Friday close)Data source: Economagic

1

2

3 B

A

1

C 2

3

A

Shock

Figure 22.4 Information shocks and underlying cycles

The important point is that a market ‘knows’ when a higher-ordertrend is beginning to develop (even if analysts don’t). Specifically, itwill not transit through certain, well-defined points unless a higher-order trend is beginning to impact. The secret is that contra-trendmovements are normally limited by their relationship to the infor-mation shocks that preceded them.

Every information shock (defined by price and energy action, not bynews flows) marks a moment in price and time where the market haschanged its perception about future potential (either in direction orextent). That moment is a point of learning. As we saw in Chapter 9,the learning process requires a diversion of energy to absorb newinformation. At the very least, this diversion of energy should notnormally make the learning organism more inefficient than it wasbefore the learning started. In the context of markets, this means thatretracements should not normally fall below the price level at whichthe information shock occurred.1 Consequently, a move back to theshock point should be taken as a warning sign.

However, there is more. The strong evidence is that markets willrelate to the shock point through the operation of the golden ratio.Specifically, the golden ratio defines the boundaries through which amarket will not go unless fundamentals are changing.

CORRECTIONS AND SHOCKS

In Chapter 10, we saw that the golden ratio – defined as 0.618:1 – isintimately bound up with the evolution of natural systems; and, in thefirst part of this book, we argued that, in effect, financial markets arenatural systems bound together by group processes. It is not, therefore,surprising to find that the golden ratio operates in financial markets. InChapter 11, we saw how derivatives of the golden ratio – especially1.618 and (1.618 × 1.618 =) 2.618 – can constrain rallies. Now weshall see how the golden ratio can constrain corrections.

The first point to make is that there are two important ‘types’ ofcorrection:

� A correction that creates the base or top pattern. For a new bulltrend, the market will make a final low, rally sharply and thenretrace a significant part of the rally. The converse holds true for abear trend. This type of correction occurs in wave 2 or wave B ofthe price pulse. It also corresponds to wave 2 of a five-wave patternor wave B in the subsequent three-wave corrective pattern.


� A correction that occurs during the impulse wave. The standardexplanation for this type is that it is a response to the marketbecoming overbought or oversold. It is therefore a ‘continuation’pattern, or period of hesitation. Broadly, it corresponds to wave Aof the price pulse in a bull run or wave 1 of the price pulse in a bearphase. More revealingly, it corresponds to wave 4 in a five-wavepattern or wave B of the subsequent three-wave corrective pattern.

The second point to make is that each trend move will be stimulatedby a very definite shock point. Such a shock point can either be thepoint of inflexion in the market, or a subsequent sharp move that isaccompanied by a jump in the energy indicators. The latter oftenoccurs as a market jumps out of a trend channel or out of a short-termtrading range (see Figure 22.5). Because of the nature of shocks, theremay be some doubt about the exact level at which it has occurred: itmay be the support or resistance level that it has penetrated; it may bethe middle of any subsequent price gap or trading range; or it mayeven be the closing price on the day of the breakout. Some degree ofleeway has to be allowed.

THE GOLDEN RATIO BOUNDARIES

The evidence is that all corrections are ‘boundaried’ in some way bythe ratio 0.618:1. That is, if the market is reversing into a new trend, itwill retrace 61.8 per cent of the initial impulse move from the appro-priate shock point to the first point of inflexion in the new trend. And


Figure 22.5 Potential shock points after a correction

if a market is only hesitating, it will retain 61.8 per cent of its wholemovement from the appropriate shock point to the point of inflexionof the current trend. As we shall see, the appropriate shock point forthe latter may be either the shock point that initiated the wholemovement or the pro-trend shock point that introduced an extensionof the trend.

If we assume, for the moment, that points of inflexion serve as therelevant shock points, then the golden ratio boundaries are as shownin Figures 22.6A and 22.6B. Hence, since 0.618 + 0.382 = 1, anymove can be divided into two parts – namely, 61.8 per cent and 38.2per cent.

IMPLICATIONS OF THE 38.2 PER CENT/61.8 PER CENT BOUNDARY

The boundary between the two proportions is of the utmost signifi-cance:

� If the boundary is accepted – that is, if it acts as support orresistance – then it will mean that the market has either learnt thata turning point has occurred or has continued to accept that funda-mentals still favour the trend.

� If the boundary is penetrated, it will mean either that the markethas not fully accepted the implications of an information shock(from a turning point) or is beginning to over-ride the implicationsof an information shock that was previously accepted.


Figure 22.6 The golden ratio

TECHNICAL CORRECTIONS ANDFUNDAMENTAL REVERSALS

There are two profound inferences here. The first is that a 61.8 per centretracement defines the maximum allowable retracement during theprocess of learning. Indeed, a 61.8 per cent retracement is part of thesignature of learning.

The second inference is that a 38.2 per cent retracement marks theboundary between a ‘technical’ correction and a more ‘fundamental’one. After receiving an information shock and moving to a new hierar-chical structure, a system must introduce constraints on its ability toregress. In the early stages after a shock, when the system is still‘learning’, there is quite a lot of flexibility. However, after the markethas accepted the implications of a shock, its ability to return to thelevel of that shock must become increasingly limited as the trendprogresses. Contra-trend moves cannot significantly influence thethrust of the main trend. Consequently, prices will respond to contra-trend ‘technical’ factors, but they will not respond to fundamentaltrends that do not exist. Hence, if prices are rising because funda-mentals are improving – and are expected to improve and, indeed, doimprove – then prices will not suddenly start to discount Armageddon.

This makes some sense. All evolving structures carry a record oftheir history in whatever it is that constitutes their ‘memory’ (that is, intheir minds and/or genes).2 They therefore ‘know’ where the last set ofboundaries has been established. In financial markets, such knowledgeis reflected in the tacit recognition of the golden ratio limitations thatare imposed on corrective movements.

TOP RETRACEMENTS

The idea that a market will retrace 61.8 per cent of the initial moveaway from a terminal juncture is a potential source of extraordinaryprofits. If, for example, there are strong grounds for supposing that themarket is peaking, then a re-test of the high is likely to retrace 61.8 percent of the fall from the top. If short positions are then opened, with astop loss,3 there is a good chance of catching the whole of the subse-quent down-wave.

Two important examples of the operation of the 61.8 per centretracement in the context of a top pattern were the US Dow JonesIndustrial Average in 1987 and the US NASDAQ in 2000. See Figures22.7 and 22.8.


On 25 August 1987, the Dow’s intra-day high was 2,746.7. A monthlater, on 22 September, the intra-day low was 2,468.99. This was a fallof 277.71 points. A 61.8 per cent retracement would have covered171.62 points, to 2,640.61. The Dow recovered to an intra-day high of2,662.4 on 2 October but closed at 2,640.99. At the close on 19October 1987, the market was 34 per cent lower.


1500

1700

1900

2100

2300

2500

2700

2900

01-M

ay-8

7

11-M

ay-8

7

21-M

ay-8

7

31-M

ay-8

7

10-J

un-8

7

20-J

un-8

7

30-J

un-8

7

10-J

ul-87

20-J

ul-87

30-J

ul-87

09-A

ug-8

7

19-A

ug-8

7

29-A

ug-8

7

08

-Se

p-8

7

18

-Se

p-8

7

28

-Se

p-8

7

08-O

ct-

87

18-O

ct-

87

28-O

ct-

87

07

-No

v-8

7

17

-No

v-8

7

27

-No

v-8

7

07

-De

c-8

7

17

-De

c-8

7

27

-De

c-8

7

Ind

ex le

vel


Source: Economagic

61.8%

retracement of

initial fall

Figure 22.7 The Crash of 1987

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

Se

p-9

8

No

v-9

8

Ja

n-9

9

Ma

r-9

9

May-9

9

Ju

l-9

9

Se

p-9

9

No

v-9

9

Ja

n-0

0

Ma

r-0

0

May-0

0

Ju

l-0

0

Se

p-0

0

No

v-0

0

Ja

n-0

1

Ma

r-0

1

May-0

1

Ju

l-0

1

Ind

ex le

vel

61.8% of

initial fall

The US NASDAQ(Weekly hi/lo/close)Source: Economagic

Figure 22.8 The NASDAQ crash of 2000–01

On 6 March 2000, the NASDAQ touched a high of 5,132.52. By 22 May 2000, it had fallen to an intra-day low of 3,042.66. This was afall of 2,089.86 points. A 61.8 per cent retracement would have been arise of 1,291.53 points to 4,334.19. In fact, the market never quite gotthere. It stopped at 4,289.06. However, this was only 1 per cent shortof target, and well within the range of error. At the close on 2 April2001, the NASDAQ was 38 per cent lower.

BASE RETRACEMENTS

A similar analysis applies to base patterns. If there are good reasons tosuppose that a base pattern might be in progress, then buying stock at(or around) the 61.8 per cent retracement level, with an appropriatestop,4 should yield very good returns.

Good examples of this phenomenon are shown in Figure 22.9. Thechart is an index of US 10-year Treasury bonds. The index is calculatedby deducting the constant maturity yield (as calculated by the FederalReserve Board) from 100. The chart shows that, at the beginning of thegreat disinflationary bull market in bonds that started in the early1980s, the base pattern was defined by the golden ratio. The bondmarket had an initial bounce off its secular low that lasted from earlyOctober to late December in 1981. This move was retraced by 61.8 per


84

85

86

87

88

89

90

91

92

93

94

Ja

n-8

0

Apr-

80

Jul-80

Oct-

80

Ja

n-8

1

Apr-

81

Jul-81

Oct-

81

Ja

n-8

2

Apr-

82

Jul-82

Oct-

82

Ja

n-8

3

Apr-

83

Jul-83

Oct-

83

Ja

n-8

4

Apr-

84

Jul-84

Oct-

84

Ja

n-8

5

Apr-

85

Jul-85

Oct-

85

Ja

n-8

6

Apr-

86

Jul-86

Oct-

86

Ja

n-8

7

Apr-

87

Jul-87

Oct-

87

Ind

ex le

vel



61.8% of initial rally = 85.30Actual low

= 85.16

61.8% of 1981–83rally = 86.40Actual low

= 86.14

Figure 22.9 The golden ratio in base patterns

cent before the market started to rally again. Overall, the first impulsewave lasted from October 1981 to May 1983. Between May 1983 andJune 1984, bonds retraced 61.8 per cent of this impulse wave.

THE GOLDEN RATIO AND CONTINUATIONPATTERNS

Once a market has started to trend, it will normally retain 61.8 per centof the move, unless underlying conditions are altering. Another wayof putting this is to say that the market will not retrace more than 38.2per cent of the trend unless conditions are altering. This is because 1 –0.618 = 0.382. Furthermore, the ratio 0.382:0.618 is equivalent to theratio 0.618:1.

Good examples of this phenomenon are shown in Figures 22.10 and22.11. The former shows the correction in 10-year US bonds in therun-up to the 1987 Equity Crash. The latter shows the fall in the DowJones Industrial Average index during that crash.

Figure 22.10 is part of the sequel to the story introduced in Figure22.9. Between October 1981 and the end of 1986, 10-year US bondsrallied by 3.30 index points. A 38.2 per cent retracement of that rallywas equivalent to a level of 89.67 on the chart. The subsequent fall in10-year bonds took the index to 89.89 on 16 October 1987.


Figure 22.10 A 38.2 per cent correction in US Treasury bonds

84

85

86

87

88

89

90

91

92

93

94

Ja

n-8

0

Apr-

80

Jul-80

Oct-

80

Ja

n-8

1

Apr-

81

Jul-81

Oct-

81

Ja

n-8

2

Apr-

82

Jul-82

Oct-

82

Ja

n-8

3

Apr-

83

Jul-83

Oct-

83

Ja

n-8

4

Apr-

84

Jul-84

Oct-

84

Ja

n-8

5

Apr-

85

Jul-85

Oct-

85

Ja

n-8

6

Apr-

86

Jul-86

Oct-

86

Ja

n-8

7

Apr-

87

Jul-87

Oct-

87

Ind

ex le

vel



38.2% of 1981–86rally = 89.67Actual low

= 89.89

We have already described the top pattern in the Dow that evolvedin the summer of 1987. On Friday 16 October 1987, equities had ashock fall. The following Monday, 19 October, the Dow plunged byalmost 23 per cent from close to close. There was further weakness thefollowing day, but the market then started to rally. It closed higher on20 October than it did on the 19th. Now, a fall of 38.2 per cent inabsolute terms from the intra-day high on 25 August (of 2,746.7)would have taken the Dow to 1,697.46. In the event the intra-day lowon the Dow on 20 October 1987 was 1,616.21. Given the trauma of theevent (and the technical difficulties in calculating the correct level ofthe index), it is not surprising that there was an overshoot. However,this was tiny in relation to the extent of the fall. Further, the marketnever closed below the 38.2 per cent retracement level. The lowestclose was 1,738.74 on 19 October.

The Equity Crash of 1987 was a traumatic event for market partici-pants. Nevertheless, taken in context, it was actually just a continu-ation pattern. The market somehow recognized the existence of animpenetrable boundary 38.2 per cent below the August 1987 high. Theimportance of this boundary – and the conclusions to be drawn from it– cannot be emphasized enough. Once the market had started to pene-trate beyond it, the implication would have been that very long-termfundamentals were deteriorating. In short, capitalism itself wouldhave been at risk.5 The point is that the 38.2 per cent/61.8 per centboundary was calculated with reference to the impulse movementsince the market was theoretically priced at zero.6 The crash event of


1500

1700

1900

2100

2300

2500

2700

2900

01-M

ay-8

7

11-M

ay-8

7

21-M

ay-8

7

31-M

ay-8

7

10-J

un-8

7

20-J

un-8

7

30-J

un-8

7

10-J

ul-87

20-J

ul-87

30-J

ul-87

09-A

ug-8

7

19-A

ug-8

7

29-A

ug-8

7

08

-Se

p-8

7

18

-Se

p-8

7

28

-Se

p-8

7

08-O

ct-

87

18-O

ct-

87

28-O

ct-

87

07

-No

v-8

7

17

-No

v-8

7

27

-No

v-8

7

07

-De

c-8

7

17

-De

c-8

7

27

-De

c-8

7

Ind

ex le

vel


Source: Economagic 38.2% fall

Figure 22.11 The 38.2 per cent fall in the Dow, 1987

1987 is one of the strongest pieces of evidence for the validity of the38.2 per cent/61.8 per cent boundary.

HIERARCHICAL STRUCTURING

It is worth noting that the golden ratio boundary is hierarchical. In abull or bear phase, the market can, in theory, retrace 38.2 per cent ofthe move to date, extend into new ground and then retrace 38.2 percent of the extended move; and so on. This process sets up a series ofturning points in a market, which will subsequently operate as supportor resistance. In short-term trading, it is easy to see why this happens.Some segments of the market will have been wrong-footed by areversal and will use the retracement to close off trades at a minimumloss. In longer-term investment, however, the reason for the operationof support and resistance is more opaque. It is certainly not the re-appearance of ancient fund managers who have been sitting on alosing trade for 20 years and who finally see an opportunity to get outat cost! At the simplest level, it is arguable that the market‘remembers’ its boundaries. At another level, however, these bound-aries are likely to implicate the golden ratio in some way.

If we assume that a market has had a series of rallies and retrace-ments, knowledge of the golden ratio presents an opportunity tocalculate future potential turning points. If the market is likely to use arecent low as support on a 38.2 per cent retracement, then the impli-cation is that the forthcoming rally into new highs will be 38.2 per centof the whole rally since the start of the move (see Figure 22.12).Hence, if a market has been trending from a low point at P1 to aninterim high at P2, and has finished a correction back down to P3, thenthe target high at Pt is:

Pt = (P3 – (0.382 P1))/0.618

THE 38.2 PER CENT/61.8 PER CENTBOUNDARY IN A DOWN-WAVE

The boundary also operates in relation to falling markets. Figure 22.13shows the performance of the US dollar against the Swiss francbetween 1978 and 2002. There was a massive rally in the dollarbetween October 1978 and March 1985. Then there was a disruptive


fall, which lasted 10 years. The most important boundary in relation tothis fall was a 38.2 per cent retracement, which was equivalent toSF1.8030. The dollar took five and a half years to get there. It wentthrough it briefly, to SF1.8298, in October 2000, and then pulledsharply back. It then rallied again in the first half of 2001, reachingSF1.8223 in the July. On neither of these two occasions, however, didthe dollar close above SF1.8030 on a monthly closing basis. It subse-quently fell quite sharply (see Figure 22.13).

GUIDELINES FOR CALCULATINGBOUNDARIES

Quite obviously, there are different golden ratio boundaries dependingon the trend that is being reversed. There are, therefore, three guide-lines to take into account:

� take note of long-term trends;

� use the five-wave ‘count’ to define a trend;

� use the concept of shock points to make the calculations.


Figure 22.12 Price projections from golden ratio retracements


The starting point is that the longer the trend has been in place, themore impenetrable will be the associated boundary – or the morework that a market will have to do to get through it. It is thereforeimportant for the analysts to take account of the history of a market.

Next, a trend can be partly defined by the basic five-wave patternthat was discussed in Chapters 20 and 21. Since the five-wave patternis basically a nested phenomenon, it is usually a simple matter to lookat a chart and isolate the beginning of each five-wave movement. Aschematic example is shown in Figure 22.14. Hence the firstboundary to be calculated is that applicable to the last wave (wave vin the dia-gram); then the bigger wave of which that is a part (wave 5in the diagram); and so on. Eventually the calculations will includethe boundaries for the whole move from the lowest low and for themove since the price was zero.

Finally the boundaries need to be calculated in relation to the actualshock points that are relevant to the trend. This is because it is theshock point that marks the level and moment in time where the marketrecognizes that the trend has been initiated or renewed. Sometimesthis is the reversal point itself. As already mentioned, this is some-times the break from a trend channel or trading range. Obviously, thelatter occur after the reversal point. Consequently, the calculatedretracement boundaries are going to be slightly higher in an uptrendor slightly lower in a downtrend. A schematic example is shown inFigure 22.15.

Figure 22.13 The US dollar, 1978–2002

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

3

Ja

n-7

8

Ja

n-7

9

Ja

n-8

0

Ja

n-8

1

Ja

n-8

2

Ja

n-8

3

Ja

n-8

4

Ja

n-8

5

Ja

n-8

6

Ja

n-8

7

Ja

n-8

8

Ja

n-8

9

Ja

n-9

0

Ja

n-9

1

Ja

n-9

2

Ja

n-9

3

Ja

n-9

4

Ja

n-9

5

Ja

n-9

6

Ja

n-9

7

Ja

n-9

8

Ja

n-9

9

Ja

n-0

0

Ja

n-0

1

Ja

n-0

2

Exch

ang

e ra

te

38.2% of 85–95 bear = SF1.8030

Swiss Francs per US$(Monthly, daily average)

Source: Bloomberg


Figure 22.14 Calculating boundaries

Figure 22.15 Calculating boundaries from shock points


PRO-TREND SHOCKS

There is another important aspect to this process of establishingboundaries. After a trend has been properly established, there maywell be a pro-trend shock that alters the deep structure of the market.In a bull market, this is a second-order change; in a bear market, it is arecognition that a longer-term reversal is under way. These pro-trendshocks can occur relatively late in a trend. Indeed, as alreadydiscussed in Chapter 9, they are likely to occur towards the location ofwhat might otherwise have been a cyclical peak. There are then twoconsequences:

� the onset of the correction will be delayed; and

� the pattern will have a five-wave structure.

There is, however, another point. This is that the pro-trend shockprovides the market with extra information that needs to be absorbedand – like any learning process – this absorption will necessitate aretracement. As has already been discussed, however, retracementsthat are associated with learning have a maximum depth of 61.8 percent of the move from the shock point to the subsequent inflexionpoint. This is demonstrated in Figure 22.16. Here, the correction willamount to 61.8 per cent of the move from the shock point to the top ofwave 3 in the five-wave pattern. The calculation also applies in thecontext of a downtrend.

The implication is that the (delayed) correction after a pro-trendshock will constitute wave 4 in a five-wave impulse movement. Thismeans, in turn, that there is another consideration. From the market’spoint of view, a wave 4 correction can only be ‘technical’. It cannotlock into fundamentals. Consequently, the correction will also need toretain 61.8 per cent of the whole move from the market low. There isthus a good chance that the market will find support at a level thatrepresents 61.8 per cent of the move after the pro-trend shock and 38.2per cent of the whole move from the low (see Figure 22.16 again). Ifthese two calculations coincide, they will provide almost impene-trable support (in a bull market) and resistance (in a bear phase).

A PRACTICAL EXAMPLE

We can now pull these strands of thought together in the form of apractical example. The market chosen is that for the US dollar in terms


of the Japanese yen. The overall profile of the dollar, since its low in1999, is shown in Figure 22.17.

To place the analysis in context, there are a number of points thatcan be made about the overall profile:

� The move since the low in November 1999 consists of five clearwaves upwards, interspersed with two corrections downwards.

� The first correction took the form of a triangle; the second took theform of a zigzag.

� Wave 3 was approximately 261.8 per cent of the widest part ofwave 2; wave 5 was 161.8 per cent of the depth of wave 4. Wediscussed the former in Chapter 11 and the latter in Chapter 20.

THE BASE PATTERN

The base formation is shown in Figure 22.18. The dollar received abullish shock, and gave a buy signal, at SP1 (ie, about ¥102.50) as itmoved out of the downtrend and cleared the trading range that had

Figure 22.16 Retracements after a pro-trend shock


established itself in December 1999. The target was 38.2 per cent ofthe last fall, which was ¥110.27. The dollar touched an intra-day highof ¥111.73 in late January 2001. Such small overshoots are notunusual in the first rally off a low.

A 61.8 per cent retracement was likely to follow. Thus the downsidetarget was either 61.8 per cent from the low or from the shock point,SP1. The targets were thus ¥105.25 and ¥106.02. In the event thedollar fell sharply, to ¥102.06, which was well below the target.

Nevertheless, that low held and the market started to oscillatewithin a triangle. The last wave of this triangle reached a low at¥104.77 in early September 2000. More to the point, the weeklyclosing low in this last stage was ¥106.00. The close is always themost critical price, since it marks a genuine point of balance. Itconformed to the principle that learning retraces 61.8 per cent of themove following an information shock.

THE WAVE 4 CORRECTION

One of the advantages of a triangle is that it gives a very clear signalwhen the market finally breaks out of it. In the system proposed byElliott, triangles are fourth-wave, or B-wave, phenomena. However,this is by no means an inviolable rule. The dollar–yen rate bears thisout. The break was not immediate. An attempt was made in October

Figure 22.17 The US dollar in terms of Japanese yen

100

105

110

115

120

125

130

135

140

145

150

Ap

r-9

8

Ju

n-9

8

Au

g-9

8

Oct-

98

De

c-9

8

Fe

b-9

9

Ap

r-9

9

Ju

n-9

9

Au

g-9

9

Oct-

99

De

c-9

9

Fe

b-0

0

Ap

r-0

0

Ju

n-0

0

Au

g-0

0

Oct-

00

De

c-0

0

Fe

b-0

1

Ap

r-0

1

Ju

n-0

1

Au

g-0

1

Oct-

01

De

c-0

1

Fe

b-0

2

Ap

r-0

2

Ju

n-0

2


Source: Bloomberg

261.8% of base = 127.35

261.8% of top = 111.08

161.8% of correction = 133.64

14

3

2

5

38.2% of the May–Nov 99

fall

Au

g-0

2

2000, which failed. However, this created waves i and ii of what wasto become the main impulse wave (see Figure 22.19). The breakoutand the buy signal occurred in late November 2000 at about ¥109.10.There was then a very sharp impulsive rally. As already described,this rally was approximately 261.8 per cent of the widest part of thewave 2 base. Once it had been completed, however, there were threedownside targets. The first applied to wave 3 itself. This was 38.2 percent of wave 3, which was ¥118.36. This was hit almost precisely inlate May. However, there were two other important downside targets.These were 38.2 per cent of the whole rally (ie, from 0 to the top ofwave 3) and 61.8 per cent of the rally from the shock point (SP2). Theformer was ¥117.06; the latter was ¥116.15. The virtual synchrony ofthe two calculations was very powerful. Having had an interim rally,the dollar turned down again. Its weekly closing low in mid-September 2001 was ¥116.58.

THE DOLLAR–YEN BEAR MARKET

The dollar then moved to a new high, thereby completing a five-wavepattern (see Figure 22.20). The buy signal was given on a break of thedowntrend in the last phase of wave 4. This occurred at ¥119.68, andis denoted SP3 on the chart. The extent of the subsequent move was,as already mentioned, approximately 161.8 per cent of the depth of


100

105

110

115

120

125

130

Apr-

99

May-9

9

Jun-9

9

Jul-99

Aug-9

9

Se

p-9

9

Oct-

99

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Apr-

00

May-0

0

Jun-0

0

Jul-00

Aug-0

0


Source: Bloomberg

SP1

38.2% of May to Nov 99 fall =

110.27

1

2

61.8% of wave 1 =

106.02

0

Sep-0

0

Figure 22.18 The dollar–yen base pattern, 1999–2000

wave 4. The next important question was: when would the bearmarket be deemed to have begun?

The first potential signal was at a level equivalent to 38.2 per cent ofwave 5, which was ¥127.76. The market went through this briefly (to¥126.40) in March 2002, but closed the week above it. The real clue,however, was the fact that 61.8 per cent of the rally since SP3 was¥125.59. In other words, the March support was probably indicatingthe importance of the shock point that triggered wave 5.

Once a market reverses through a 61.8 per cent retracement from ashock point, it has basically rejected the learning that was triggered bythat shock point. In other words, once the levels ¥127.76 and ¥126.40were broken, the dollar was in a bear phase. And so it was. The marketsubsequently fell towards a level that was 61.8 per cent of the whole1999–2002 rally.

CONCLUSION

This chapter has explored some of the relationships between infor-mation shocks and the golden ratio. It was hypothesized that thegolden ratio is intrinsic to the learning process, such that markets areunlikely to retrace more than 61.8 per cent of any rally after an infor-mation shock. It was also hypothesized that markets would retain 61.8per cent of their move since an information shock once learning had


100

105

110

115

120

125

130

Se

p-9

9

Oct-

99

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Apr-

00

May-0

0

Jun-0

0

Jul-00

Aug-0

0

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Apr-

01

May-0

1

Jun-0

1

Jul-01

Aug-0

1

Se

p-0

1


Source: Bloomberg

SP2

1

2

38.2% of whole rally= 117.06

61.8% of SP2 to 3= 116.15

0

i

ii

iii

iv

v

3

38.2% of wave 3

= 118.36

4

Figure 22.19 The wave 4 correction, April–September 2001

occurred. A break of either of these two boundaries would likelysignal a change in deeper-running fundamentals.

It was demonstrated that this hypothesis seems to work in practice.There is obviously some scope for error in calculating the appropriateboundary levels because fast-moving markets will rarely hit themprecisely. On the other hand, the markets will not signal that they havebroken a boundary, unless deeper-running fundamentals have indeedstarted to reverse.

NOTES

1. So if the information shock is accompanied by a price gap in the market, thisgap should not normally be filled. This contradicts the old stock market adagethat ‘gaps are always filled’.

2. Jantsch, Erich (1980) The Self-organizing Universe, Pergamon, Oxford.3. The stop will depend on the user’s preferences. One is to place the stop just

beyond the turning point. This was the stop favoured by Gartley, who actuallymade a great deal of money from selling the simple rule as a ‘system’ duringthe Depression years. See Gartley, H M (1935) Profits in the Stock Market,reprinted (1981) by Lambert-Gann, Pomeroy, Washington. However, this maybe too far away from the entry point. Another is to place the stop (eg) 1 per centaway from the entry point. If the stop is triggered, the market may just be over-shooting, so positions should be re-entered as the market drops back below the61.8 per cent level.

4. See note 3.


100

105

110

115

120

125

130

135

140

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Apr-

01

May-0

1

Jun-0

1

Jul-01

Aug-0

1

Se

p-0

1

Oct-

01

No

v-0

1

De

c-0

1

Ja

n-0

2

Fe

b-0

2

Ma

r-0

2

Apr-

02

May-0

2

Jun-0

2

Jul-02

Aug-0

2


Source: Bloomberg

SP3

2

61.8% of whole rally= 114.19

3

61.8% of rally since SP3 = 125.594

538.2% of wave 5 =

127.76

Figure 22.20 The bear phase in dollar–yen, 2002

5. This does not rule out (and did not rule out) the possibility of a subsequentrecession. There was, in fact, a deep recession in 1990–92. The combination ofa crash and deep recession tends to occur in the first Juglar cycle after aKondratyev commodity price high (see Chapter 17).

6. The same conclusion applies to the low in the Dow that was reached in earlyOctober 2002. A 38.2 per cent fall from the all-time high of 11,750.28 was7,260. The market’s low on 9 October was 7,197.45, after which the marketbounced. It remains to be seen whether this level will eventually be penetrated.


326

23

The confirmation of buyand sell signals

INTRODUCTION

The greatest constraint on taking the appropriate action in any financialmarket is doubt. No matter how important a trading signal mightappear to be, there is always the possibility that it might be wrong. Asimportant buy and sell signals almost inevitably occur at momentswhen the vast majority of investors are literally ‘the other way’, thecrowd pressure to ignore the signal or wait for a more certain one islikely to be extremely strong. The first task of any investor, therefore, isto try to reduce their susceptibility to such crowd pressures.

In this context, people who use technical analysis to trade in financialmarkets will quickly find that their ability to stand out against the herd isincreased. This is quite simply because the raison d’être of technicalanalysis is to deduce what the vast majority are indeed doing, so thatcontrary positions can be taken at the appropriate moment.Nevertheless, it is always essential to try to reduce the element of doubtby applying as many techniques as possible to each particular situation.The interpretation of price patterns is only one of these techniques.Others involve the calculation of price objectives, the calculation of thetiming of likely turning points and a direct analysis of investorsentiment and behaviour. In this chapter, we shall examine the idea thatwe can deduce the exact position of the market in its bull–bear cycle byanalysing investor behaviour.

The confirmation of buy and sell signals 327

INVESTOR CONFIDENCE ANDPRICE FLUCTUATIONS

The starting point to the analysis is the relationship between investorsentiment or investor activity and price fluctuations. It will be remem-bered that we examined the relationship in some detail in Chapter 8.There it was argued that there is both a limit cycle relationship and aspiral relationship between investor confidence (measured bysentiment or by behaviour) and prices. While a trend is intact, changesin confidence encourage changes in prices, which in turn stimulatechanges in confidence and investor activity. However, at a turningpoint, the following sequence of events occurs:

� the ‘circular’ relationship between prices and confidence begins tobreak down, changing prices are unable to induce additionalchanges in investment positions and, accordingly, the marketbecomes overextended;

� investor psychology receives a shock as prices move in theopposite direction to the expectations of the vast majority;

� prices move to re-test the levels that were reached just prior to theshock (sometimes prices actually move into new territory duringthe re-test; sometimes they do not, but, during this phase, confi-dence in the price trend often appears to run ahead of the actualchange in prices, although it does not genuinely return to levelsachieved prior to the shock);

� finally, prices begin a proper reversal and investor confidenceswitches with it.

OVEREXTENDED MARKETS AND THEPRINCIPLE OF NON-CONFIRMATION

This analytical framework not only allows investors to understandexactly what is happening at each phase of a full bull–bear cycle, butit also pinpoints exact trading rules that can be used to take advantageof this understanding:

� it emphasizes that it is possible to anticipate an energy gap – and,hence, an information shock – when the market becomes overex-tended in one direction or another;

� it highlights the fact that investor psychology becomes damagedby the shock, so re-tests of the price levels that were reached priorto the shock can be used to open up new trading positions in theopposite direction.

In practice, then, the main indicator of an imminent price reversal is achange in the ‘level’ of any index that measures either investorsentiment or investor behaviour. When a trend is intact, the majorityof investors will open up new trading positions, increase their dealingactivity, be prepared to stimulate the momentum of the price trend anddeal in all sectors of the market. However, when a price reversal isimminent, the relevant confidence indicators tend first to becomeoverextended, then drop away because investors are simply unable toopen new positions. If the market does then fall, the critical question ishow the confidence indicators respond during subsequent recoveryfrom the shock. If there is to be a third stage to the reversal, theninvestors will be less willing than previously to open up new tradingpositions. And if prices actually move out into new territory, but theconfidence indicators do not, then the trend is not confirmed.

These relationships are demonstrated in Figure 23.1. In effect, ahigher-level energy gap is signalled by the performance of the confi-dence indicators during the fourth and fifth waves of a distortedlower-level price pulse.

INDICATORS OF INVESTOR BEHAVIOURIt is therefore important to be able to gauge the disposition of investorsentiment and the strength of investor activity. The most obvious indi-cators to use are opinion surveys and liquidity surveys. However,these can be expensive to create or to buy, and are usually late becausethey take time to calculate. A much easier approach is to use simplemechanical indicators that are readily available from the marketsthemselves. These include:

� price momentum (for example, percentage rates of change);

� price volatility (for example, percentage spread intra-day or intra-week);

� trading volumes (that is, number of bargains or nominal turnover);

� open interest and put/call ratios (ie, outstanding positions); and

� related market price indices that have a different focus.


These do not exhaust the list, but they have the advantage of beingeither easily calculable or easily available. The first two can be calcu-lated directly from prices themselves, and the rest are usuallyavailable from the relevant bourses.

VOLUME AND OPEN INTEREST

Because the rules applying to volume and open interest are essentiallythe same, we shall analyse them together. Volume is a direct measureof the amount of activity taking place in the market at a particulartime. It may be measured either in terms of nominal turnover or thenumber of bargains. The data may be obtained from the cash markets,futures markets or even the options markets. Indeed, in principle thevolume figures for one market for a particular asset may be applied toany of the other markets in the same asset: all that is required is thatthe data be consistent and the markets be well traded. Hence, for


Figure 23.1 The principle of non-confirmation

example, in the foreign exchange markets where it is impossible toobtain volume data, it is perfectly acceptable to use volume figuresgenerated by the relevant exchange for an equivalent futures contract.

Open interest, on the other hand, is the total number of outstandingcontracts in either the futures or options market that are held bymarket participants at the end of each day. Although the measureapplies to specific futures or options markets, it may also be used forthe related securities.1 Hence, for example, the open interest for theUS Treasury long bond futures contract can be used in conjunctionwith the Treasury cash market.

THE LEVEL OF VOLUME

There are two aspects of open interest and volume that warrantdiscussion. The first is the general level of the indicators, while thesecond is the direction of change of those indicators. Let us start withthe levels. The level of volume is indicative of people’s willingness todeal. This, in turn, reflects traders’ attitudes to the market. A low levelof volume indicates an unwillingness to open new positions and closeold ones. It therefore also indicates some uncertainty about the futuredirection of the market. Alternatively, a high level of volume is adirect reflection of traders’ willingness to open new positions, takeprofits or close bad positions. It therefore also indicates a high degreeof confidence in the future direction of the market.

THE LEVEL OF OPEN INTEREST

In the case of open interest, the level is indicative of the efficiency, orliquidity, of the market. Specifically, it is a measure of the market’sability to absorb new deals at current prices. If open interest is low, thenthere are very few profits to be taken or bad positions to be closed. Anew trade is likely to have a large impact on prices because prices haveto adjust in order to induce another dealer to ‘go the other way’. At lowlevels of open interest, therefore, the market is illiquid. If, on the otherhand, open interest is high, then there are plenty of profits to be takenand bad positions to be closed. A new trade is likely to have only a verysmall effect on prices because another dealer is likely to take advantageof the availability of the trade at the current price level. At high levelsof open interest, therefore, the market is very liquid.


SUDDEN CHANGES IN THE INDICATORS

The concepts of ‘high’ and ‘low’ are, of course, relative. There needsto be a benchmark against which to measure them. In this respect,historical precedent and local market knowledge should provide someassistance. Very often, however, the most important facet of the levelsof either open interest or volume is the fact that they change suddenly.This is a clear indication that something is altering in theprice–sentiment relationship. In this context, volume is often moreimportant than open interest, but open interest can help to determinethe extent of the price reversal that subsequently takes place. Let ustherefore analyse the implications of the direction of change involume and open interest in more detail.

THE DIRECTION OF CHANGE IN VOLUMEAND OPEN INTEREST

Both volume and open interest may be taken as proxies for investorconfidence in the direction of the prevailing trend. Rising volumesuggests a growing awareness of the higher-level trend, and risingopen interest indicates a growing commitment to that trend. Fallingvolume, on the other hand, indicates a spreading inability or unwill-ingness to pursue the immediate trend, while falling open interestsuggests some reversal of sentiment as profits are taken and bad posi-tions are closed out.

CHANGING EMOTIONS DURING THE CYCLE

It will be remembered that while a trend is developing, the limit cycleprocess between prices and confidence will ensure that investoractivity continually strengthens that trend. This implies that, duringthe trend, volume and open interest should both rise. This occurswhether the price trend is upwards or downwards. At some stage in thelimit cycle process, the change in prices will trigger the emotion offear among investors – either the fear of being left out of the marketduring a rise or of being left in the market during a fall. Most emotionshave a number of different dimensions, but fear is unique in closingthe mind to rational thought and focusing the body’s energy on


physical, mental or social survival. Fearful investors will deal blindlyand both volume and open interest should rise sharply. Consequently,the market starts to become overbought or oversold.

SHARP RISES IN VOLUME ANDOPEN INTEREST

A sharp rise in volume and open interest essentially undermines thelimit cycle between confidence and prices in the short term. Inparticular, the rise in open interest creates conditions where investorsare overstretched and are easily induced to close their positions inresponse to relatively small changes in price. On the one hand,therefore, investors are unable to marshal sufficient resources toperpetuate the old trend, while, on the other, profit-taking or bearclosing may emerge quickly as prices start to reverse. Consequently,the reversal will become self-generating.

THE REVERSAL PROCESS

This is, however, only part of the process. Notice what happens as thereversal gathers pace. Investors close profitable positions and cut outlosing ones. Not all of them will react, of course, but many will. Thismeans that, during the first stage of a reversal, open interest will eithernot rise or will actually fall. At the very least, therefore, the immediateprice trend is not being confirmed. At some stage, therefore, priceswill begin moving in the same direction as the old trend again and willre-test the levels at which the market became overbought or oversold.It is this re-test that will provide the true indication of the futuredirection for the market.

VOLUME AND OPEN INTEREST DURINGFIFTH WAVES

As we have already seen, a re-test may actually move into new territory.This new high or low is counted as a ‘fifth wave’, caused by a strongunderlying trend. There are four immediate possibilities with respect tothe associated movements in volume and open interest:


� volume and open interest rise into new ground;

� volume and open interest rise, but not into new high ground;

� volume rises, but open interest falls;

� volume and open interest fall.

It follows from the nature of the price–sentiment cycle that if volumeand open interest both respond strongly to a new high or low, then thehigher-level trend remains intact. Consequently, the damage inflictedby the earlier overbought or oversold condition will have beenovercome and the price reversal will have been no more thantemporary.

If, however, volume and/or open interest do not manage to rise intonew high ground, then a non-confirmation takes place. (This wasshown in Figure 23.1.) This means that the damage inflicted oninvestors’ psychology by the initial price reversal was serious and animportant trend reversal is about to emerge.

This last conclusion is even more apt if open interest falls during avolume non-confirmation. The implication here is that traders aretaking advantage of new price levels to close positions. Confidence inthe old trend has therefore already altered and the subsequent pricereversal will be strong.

Finally, and in potentially the worst case of all, if prices move intonew ground, but volume and open interest fall, then the subsequentprice reversal could be quite dramatic. The limit cycle has obviouslyalready broken down completely and investors are in a position torespond with vigour when a new price trend emerges.

VOLUME AND OPEN INTERESTDURING RE-TESTS

Remember that an ordinary, undistorted, re-test of a previous peak iscarried out by a wave B, and the re-test of a previous low is carried outby a wave 2. By definition, sentiment during either of these waves islikely to fall short of that which occurred during the precedingimpulse waves.2 This internally weak technical position of the marketcompounds the likely implications for the upcoming impulse wave.There are then actually three relevant combinations of volume andopen interest:


� volume and open interest rise;� volume rises, but open interest falls, or vice versa;� volume and open interest fall.Volume and open interest may well rise together as the sentiment of theprevious trend reasserts itself. Usually, however, one or both remainsubdued and will not reach the levels attained during that trend.3

The real clue to future price movements may, in fact, be given by theperformance of open interest. If open interest falls, the evidencesuggests that investors are closing positions. This is, of course, truewhether volume is rising or falling.

If investors are closing positions, then the implication is that thedamage done by the previous shock to the market has not beenovercome. The greater is the fall in open interest, the less the market issubsequently able to cope with changes in the supply of, or demandfor, stock, and consequently, when an impulse wave materializes, theeffect will be dramatic.

THE WIDER IMPLICATIONS OF FALLINGOPEN INTEREST

The foregoing analysis highlights the fact that falling open interest(particularly if it is accompanied by rising volume) can provide a veryaccurate warning of an impending ‘third’-wave reversal in prices. Infact, falling open interest can also warn of an impending accelerationin prices just after a price reversal. It is as well to be clear about thedifference between the two situations.

Remember that falling open interest essentially restricts the flexi-bility of the market. The subsequent move is likely to be large becausesmall changes in demand or supply have a large effect on prices, and


Table 23.1 The bull–bear cycle and open interest

Prices Open interest

1. Bear squeeze Sharp fall2. Re-test of low/new low Rise, but not to new high3. Beginning of bull Fall4. Bull Rise5. Overbought set-back Sharp fall6. Re-test of high/new high Rise, but not to new high7. Beginning of bear Fall8. Bear Rise

large changes in prices then induce a significant change in sentiment,so the process becomes self-generating. Table 23.1 shows the likelyfluctuations in open interest during the course of a theoreticalbull–bear cycle. Put bluntly, falling open interest occurs when the‘winners’ are squeezing the ‘losers’. Hence, generally speaking, asharp fall in open interest marks the ‘death’ stage of the life cycle ofeither a bull market or a bear market. It therefore marks the beginningof a significant reversal.

Alternatively, a moderate fall in open interest may mark the ‘birth’stage of a new bull or bear trend. It therefore often precedes an accel-eration into the main thrust of that trend. Figure 23.2 shows thesepossibilities (without taking account of high-level trends) within thecontext of a price pulse.


Figure 23.2 The price pulse and open interest

MOMENTUM ANDOVEREXTENDED MARKETS

The third analytical tool that can be used to judge the internal strengthof a market is a momentum index. Essentially, such an index is ameasure of the speed of change in the market. The value of using amomentum measure is actually two-fold. First, the price–sentimentlimit cycle suggests that, during an impulse wave, momentum willaccelerate. Furthermore, momentum will be at an extreme at the endof these waves – analysts use the terms ‘overbought’ or ‘oversold’ toindicate these extremes. Having reached an extreme, the marketshould reverse itself. If the price pulse is a high-level pulse, thereversal will be immediate, sharp and long-lasting.4 If, however, thepulse is of a lower order, the reversal will be only temporary.

MOMENTUM AND NON-CONFIRMATIONSecond – and if the reversal is only temporary – a momentum measurecan be used to judge the strength of the renewed trend. The principleof non-confirmation relies on the idea that the momentum of a market(whether upwards or downwards) will slow appreciably during thefinal wave (Elliott’s ‘fifth’ wave) that precedes an important reversal(see Figure 23.1 again).

Indeed, this can often be seen in the form of a change in the slopeof price movements on the price–time charts during the final wave.Momentum non-confirmations occur both because the short-termlimit cycle relationship between prices and confidence is inherentlybiased (see Chapter 8) and it is, in any case, being gradually over-ridden by longer-term forces. If higher prices are unable to encouragebuying or lower prices are unable to encourage selling, a reversalmay be imminent.

MEASURES OF MOMENTUM

There are four different measures that can be used as a momentumindex:

� a simple percentage rate of change, where the current price isexpressed as a percentage of the value of an earlier price – the


‘earlier’price is always a constant time period (hours, days, weeks,months and so on) away from the ‘current’ price;

� the deviation from a (long-term) moving average – the differencebetween the current price and the moving average (the movingaverage has a constant number of data points in it and includes thecurrent price as the last such data point);

� a relative strength indicator (RSI), which measures the rela-tionship between price increases and price decreases;

� a directional indicator, which directly measures the strength of atrend.

RATES OF CHANGE

Each of the different momentum measures has its own advantagesand disadvantages. The percentage rate of change index, for example,is certainly very easy to calculate. Furthermore, it is possible both touse indicators with very short time periods (such as five days) toisolate the momentum of lower degree price pulses and measureswith long time periods (such as 12 months) to analyse higher-degreepulses. In order to obtain some idea of what constitutes overboughtand oversold, however, it is necessary to have a long historyavailable. Extremes of momentum can then easily be determined byinspection (see Figure 23.3).


Mom

entu

min

dica

tor

Figure 23.3 Overbought and oversold momentum

DEVIATIONS FROM A MOVING AVERAGE

The use of deviations from a moving average is almost as popularbecause moving averages are a proxy for the market trend. The theorybehind using the deviation from the moving average implicitly relieson the fact that lower-degree price pulses cannot deviate too dramati-cally from the trend imposed by higher-degree pulses. Consequently,it is possible to test the historic data to find out what constitutes over-bought or oversold. Furthermore, a narrowing of the deviation whenthe market reaches new highs or lows is taken to mean a reduction inthe power of the lower-degree pulse vis-à-vis the (gradually reversing)higher-degree pulse.

Essentially, there are two possible systems that use moving averages.The first measures the deviation of the current price from an arithmeticalmoving average of recent prices. This method is very simple, but is notwithout some imperfections. Specifically, the calculated value of anarithmetical moving average applies to the data at the mid-point of thetime period over which it is calculated. Consequently, it is not possible toknow the ‘correct’ value of a moving average that is applicable to todayuntil some time in the future. The system actually works partly becauseof the idiosyncrasies of the moving average technique itself. The movingaverage will lag the current price quite significantly. Hence, overboughtand oversold conditions occur simply because the moving average hasnot ‘caught up’ and non-confirmations occur simply because slowlymoving final waves actually do allow the moving average to catch up.

The second form of the moving average system addresses theseproblems by using two exponentially smoothed moving averages.Exponential smoothing gives a heavier weight to the most recent timeperiods and a lesser weight to earlier ones. It therefore measuresrecent crowd pressures, but concentrates its power on recent events.Today’s exponential moving average (EMAt) is calculated as:

EMAt = Pt × K + EMAt–1 × (1 – K)where:

K = 2/(n + 1)n = number of days in moving averagePt = today’s priceEMAt–1 = yesterday’s moving average.

Subtracting the longer-term exponential moving average from theshorter one results in an indicator that is normally referred to as a


‘moving average convergence–divergence’ oscillator (MACD). Thisindicator – originally developed by Gerald Appel in the United States –oscillates slowly as the short-term average converges on, crosses andthen diverges from, the longer-term average. The MACD has becomeone of the primary tools used to judge overbought and oversold condi-tions and signal non-confirmations.

THE RELATIVE STRENGTH INDEX (RSI)

The RSI was developed by J Wells Wilder in the United States. Itmeasures the relationship between the sum of the daily price increasesduring a given recent period (usually the last nine days) and the sum ofthe daily price decreases during the same period. The formula for themeasure is given as:

Current RSI = 100 – 100

1 + Sum of positive changesSum of negative changes

If the sum of all the positive changes during the chosen time period iszero, the ratio in the denominator is also assumed to be zero. However,if the sum of the negative changes is zero, the ratio in the denominatoris assumed to be equal to the sum of the positive changes (that is, thesum of the negative changes is assumed to be unity).

The RSI is particularly easy to use. It requires only a very shorthistory, needs no testing to find out what constitutes overbought oroversold and is easy to present graphically. The RSI can oscillate onlybetween values of 0 and 100. Generally speaking, if the value is 20 orless then the market is likely to be oversold; if the value is 80 or more,the market is likely to be overbought. The RSI is also well suited toestablishing divergences at market troughs and peaks.5

It has to be remembered, however, that the RSI is essentially ashort-term trading tool. Hence, a nine-day RSI implicitly assumes thatit is unusual for a market to move in one direction for more than eightconsecutive trading days. It is nevertheless worth checking thehistorical data to ensure that uni-directional movements of eight daysor more are indeed uncommon. If necessary, the RSI can be calculatedusing longer time periods.


MOMENTUM TRADING RULES

Whichever method is used to track momentum – whether it be rates ofchange, deviations from a moving average or the RSI – there are threeimportant rules that should be borne in mind:

� Markets can remain overbought for quite a long period of time. Aswe have already observed, fear of missing further profits (that is,greed) takes time to dissipate. Traders should not therefore reactsolely to an overbought condition unless there are other groundsfor supposing that a reaction is imminent. Errors can be avoided bywaiting until the price level and momentum have both started tofall from their overbought level.

� Markets tend to remain oversold for relatively short periods oftime, because the fear of making losses tends to trigger swiftcorrective action. Hence, purchases based on oversold criteria willusually yield a profit.6 Most importantly, non-confirmation of newprice highs or lows only occurs when prices themselves (andtherefore the momentum index) actually start to reverse.

� Many analysts make the mistake of assuming that because priceshave moved into new territory while momentum has not, a pricereversal is imminent. It is, in fact, often true that either the non-confirmation persists for an extended period of time or momentumeventually catches up with the market.

THE DIRECTIONAL INDICATOR

The fourth momentum indicator – the directional indicator – has avery specific use: it measures the strength of a trend. This is obviouslyof great value in deciding whether the market is trending or oscillatingin a trading range, and in deciding if the trend is pronounced or not.The original work on this index was (again) conducted by J WellsWilder,7 although it has been modified by other analysts. The direc-tional indicator (usually known as the DI) is based on the assumptionthat an up trend exists if today’s high exceeds yesterday’s high, whilea down trend exists if today’s low exceeds yesterday’s low. Theaverage directional indicator (known as the ADX) is based on theassumption that the more often a market generates a trading rangeoutside the previous day’s trading range and the greater are these


external trading ranges in relation to the day’s total move, then themore powerful is the trend.

The mathematics of the directional indicators is quite complicatedand is best produced by computer. Most technical analysis softwarepackages now include them. All the main indicators are ‘normalized’in as much as they are multiplied by 100. This means that the numbersoscillate within a range of 0 to 100. For those who wish to observe thecalculations directly, the details are included in Wilder’s original bookand in books such as Alexander Elder’s excellent Trading for aLiving.8 Here we shall just confine ourselves to the use of the indi-cators.

The beauty of the DI and the ADX is that, between them, theydirectly track the evolving power of the crowd. There are two direc-tional indicators – a positive one (+DI), which measures up days, anda negative one (–DI), which measures down days. The positive andnegative numbers are averaged over time. If the moving average of the+DI exceeds the moving average of the –DI, then the market is rising.In principle, a buy signal is given if the +DI line crosses the –DI linefrom below, and, again. The reverse holds true for falling markets.

Meanwhile, the ADX – which is widely regarded as the mostimportant component of Wilder’s system – measures the spreadbetween the +DI and –DI lines. In fact, it is a moving average of thespread. Hence, if a trend is evolving in a healthy manner, the ADX linewill be rising. This applies whether the trend is up or down. On theother hand, if a trend is running out of steam, the ADX line will beginto reverse. The market is becoming overextended. Importantly, too,the relevant DI line9 is likely to be signalling a non-confirmation atthis point. Finally, if a market is not actually trending, but, instead, isfluctuating horizontally, then the ADX will be falling and is likely tobe below both the +DI and the –DI.

THE ADVANCE–DECLINE INDEX

The final set of indices consists of auxiliary indices. These are used toestimate the extent to which all stocks or all sections of the market areparticipating in a trend. There are two basic auxiliary indices that canbe used. The first is an index that is constructed from the daily figuresfor the number of stocks that have advanced in relation to the numberof stocks that have declined. The most popular method of constructionis simply to add the net daily advances to, or deduct the net dailydeclines from, a cumulative figure. The resulting index is known as


the ‘cumulative advance/decline indicator’, or the A/D line.Obviously, such figures are not readily available for all markets, butthey are available for stock markets in both the UK and the UnitedStates. As they relate to the total number of stocks being traded in themarket, they readily provide a good proxy for its internal strength.Hence, if a market price index moves into new ground, but the A/Dline does not follow suit, then there is a strong suggestion that tradingactivity is becoming too speculative and that not all stocks are partici-pating in the movement. Such a non-confirmation does not necessarilymean that it is wrong to participate in the market, but it does mean thatthe life expectancy of the movement may be limited10 and that it couldbe followed by a relatively severe setback.

A SECOND PRICE INDEX

The alternative type of auxiliary index involves the use of a secondprice index. It must, however, consist of stocks that are entirelydifferent from, but nevertheless related to, those included in theprimary index. Hence, in the fixed interest markets, for example, ashort-term (interest rate-related) bond index may be used tocomplement a long bond index. In the US stock market, the Dow JonesTransportation Index may be used to complement the Dow JonesIndustrial Index; in the gold market, the price of gold itself may be usedto complement a gold mining share index. In all cases, it is reasonableto expect that if a trend is developing in a healthy fashion, both primaryand secondary indices should perform together. A divergence betweenthe two indices suggests that a reversal might be imminent.

THE PRINCIPLE OF DIRECT CONFIRMATION:THE DOW THEORY

This chapter has so far covered the basic notion that non-confir-mation provides a useful warning of an imminent reversal. The finalstage of the analysis is to see how technical indicators can be used toconfirm directly the validity of buy and sell signals based on pricepatterns.

The first technique – which involves the use of an auxiliary priceindex – was first introduced by Charles Dow.11 Dow likened move-


ments in the market to the behaviour of the sea: the primary trend in themarket (which lasts for at least a year) corresponds to the direction ofthe tide; the secondary reactions (which last for three to 13 weeks) arethe waves, and movements of lesser degree are the ripples on thewaves. The first tenet of Dow Theory is that a bull market is formed bya series of rising peaks and rising troughs, while a bear market isformed by falling peaks and falling troughs. The second tenet of thetheory is that, as two stock indices are part of the same ‘ocean’, thetidal action of one must be reflected in the tidal action of the other.

A trend is therefore considered to remain intact until a reversalsignal – based on an analysis of the peaks and troughs – is generated.In the first place, a reversal signal consists either of lower highs andthen lower lows in what has been a rising market, or of higher lowsand then higher highs in what has been a falling market. In otherwords, a reversal from a top will consist of an initial fall, a rallyupwards to re-test the high, and then a drop beneath the low of theinitial fall; while a reversal from a trough will consist of an initialrally, a setback to re-test the low, and a rally beyond the high of thefirst rally. These are shown in Figure 23.4.

At the same time, however, an apparent reversal signal will not bevalidated unless all sectors of the market are involved. Hence, asecond price index, which is separate from but related to the primaryindex, must confirm the primary index. However, there are two possi-bilities: one, where the secondary and primary indices are synchro-nized; and two, where the secondary index diverges from the primaryindex at what subsequently turns out to be the re-test of the turningpoint for the primary index. This latter case is shown in Figure 23.5.Here, the secondary index makes a new high or low as the primary


Figure 23.4 Dow Theory reversal signals

index re-tests its turning point. This is a non-confirmation. Then, thereversal signal is generated when the secondary and primary indicesboth move beyond the extremes of the preceding contra- (old) trendcorrections. At this stage, the secondary index need not have estab-lished a series of rising peaks and troughs (for a bull) or falling peaksand troughs (for a bear).

Quite obviously, these reversal patterns conform to the three-wavearchetype that has been explored in earlier chapters. Equally obvious,however, is the fact that the emergence of a three-wave pattern maycomplete a contra-trend move, rather than signal a change in trend.Dow Theory deals with this problem by paying attention to volumesand by ensuring that the waves that are being tracked are of the correctorder of magnitude. Often there will be a high-volume event just priorto the ultimate high or low, and/or during the re-test of the high or low.There should then be a pick-up in volumes as the reversal signal isgiven. In addition, for a signal to be valid, the time elapse of the initial


Figure 23.5 Non-confirmation and confirmation in Dow Theory

fall from a high, or rally from a low, must be between three and 13weeks – never shorter, and seldom longer.

By tradition, Dow Theory buy and sell signals are generated byreference to the Dow Jones Industrial Average and the Dow JonesTransportation Index. A buy signal is shown in Figure 23.6.Industrials reached the low of their massive 2000–02 bear market inOctober 2002. Transports also made a low at the same time. Bothmarkets then rallied but, when Industrials reached the end of their re-test of the low in March 2003, Transports actually fell to new lows.This was a non-confirmation. However a non-confirmation is not, initself, a signal: Industrials could, at a later stage, have followedTransports to new low ground. In the event, a buy signal was given on4 June 2003, when Industrials closed above the peak of theirNovember 2002 rally high. Transports had already risen above theirprevious contra-trend rally high on 2 May 2003.

The main criticism of Dow Theory is that the signals it provides arevery late. However, their importance lies in the fact that they areusually very accurate: an apparent reversal in only one index is oftenaborted if it is not confirmed by the second index, and a confirmedbuy or sell signal indicates that all sections of the market are partici-pating in the move.


Figure 23.6 Full Dow Theory buy signal

4500

5000

5500

6000

6500

7000

7500

8000

8500

9000

9500

10000

10500

11000

11500

Sep

-01

Nov

-01

Jan-

02

Mar

-02

May

-02

Jul-0

2

Sep

-02

Nov

-02

Jan-

03

Mar

-03

May

-03

Jul-0

3

Sep

-03

Nov

-03

Jan-

04

Mar

-04

May

-04

Jul-0

4

Sep

-04

Nov

-04

Jan-

05

Mar

-05

May

-05

Jul-0

5

Sep

-05

Nov

-05

DJ

Indu

stria

ls

1600

1800

2000

2200

2400

2600

2800

3000

3200

3400

3600

3800

4000

4200

4400

4600

4800

5000

DJ

Tran

spor

ts

Transports(r.h.scale)

Dow Jones Industrial and Transport Averages(Daily closes)


Industrials(l.h.scale)

Non-confirmation

Buy

THE PRINCIPLE OF DIRECT CONFIRMATION:OTHER INDICES

Similar conclusions can be drawn from the use of sentiment indi-cators. If volume, open interest and/or momentum start to confirm aprice movement immediately after a price reversal, the reversal islikely to be important. In particular, a sharp change in momentum (tothe extent of either becoming very overbought or very oversold veryquickly) strongly suggests that the limit cycle relating prices tosentiment has reversed.

CONCLUSION

Let us now summarize the conclusions of this important chapter. Thelimit cycle relationship between prices and sentiment postulates that aprice trend is essentially intact while confidence is improving and themajority of stocks are participating in that trend. Confidence may berepresented by a number of different analytical tools, of which the mostpopular are volume, open interest and momentum. Hence, a trend isintact while all these indicators are still increasing. Such changes aresupportive: increasing dealing activity and accelerating momentum(whether up or down) are healthy indications of profits being taken,bad positions being closed and new positions being opened. Sooner orlater, however, the confidence indicators will reach excessive levels(often in a surge) that cannot be sustained. They will therefore subse-quently tend to contract, even though prices continue to trend. Thiscontraction is strong evidence that an energy gap is appearing. It istherefore non-confirming: prices will need to undergo a contra-trendcorrection as the overbought or oversold conditions are unwound.

This contra-trend correction may either be swift and temporary orthe first stage of a more serious reversal pattern. In either case, the pricelevels associated with the original overbought or oversold conditionwill always be re-tested to some degree. If the re-test takes the marketto new highs or lows, then the original trend will remain intact unlessvolume, open interest and momentum are unable to respond. If, on theother hand, the re-test falls short of the initial price turning point, thenthe market is potentially very vulnerable to a more extended reversal.Hence, either a non-confirmed new high or low (or an unsuccessful re-test) becomes the second stage of the reversal process. The third stageis a dramatic thrust in the direction of the new trend.


These conclusions will cover the majority of situations thatinvestors will meet. There are, however, three other techniques thatcan be used to ensure against major errors:

� First, remember that volume and open interest are essentiallyshort-term indicators. It is therefore useful to track at least onelong-term indicator.

� Second, financial markets respond to the influence of the goldenratio and tend to reverse when these influences come into effect;golden ratio-defined price targets may be used to confirm theconclusions from other methods.

� Third, the impact of limit cycles implies that markets contain arhythmic beat within their oscillation mechanisms. Cycle analysesmay therefore be used to pinpoint the moments in time whenexcesses of sentiment may be expected to occur. It is therefore to theconcepts of golden ratio constraints and time cycles that we next turn.

NOTES

1. An alternative measure is the ratio of puts to calls in the options market. A highratio reflects a large percentage of open bear positions, and a low ratio reflectsa large percentage of open bull positions. The analysis of this chapter can beapplied to the put:call ratio.

2. In a normal re-test, the concept of non-confirmation does not apply.3. There is an important point here that is often missed. Volume and open interest

analysis is only applicable to relatively short-term market conditions. It ispossible, therefore, to pinpoint long-term reversals only because of an accurateinterpretation of the short-term. It is not valid to compare today’s open interestor volume with that which occurred, say, two years ago. The market structure islikely to have altered over such a time period.

4. For example, cyclical lows in the Dow Jones Industrial Averages in 1932,1942, 1949, 1957, 1962, 1966, 1970, 1974, 1978 and 1980 occurred simultane-ously with momentum lows. However, it should be added that this does notmean that the low will never be re-tested. The influence of a large cycle justmeans that the re-test occurs after such a significant time period that it cannotbe used for short-term trading.

5. When testing for peaks, the momentum high at the top of wave 3 should beachieved with an RSI of 80 or more. The market will then correct for the over-bought condition and the RSI will fall. Prices and the RSI will both subse-quently rise again as wave B (the Elliott fifth wave) comes into force. Priceswill move into new high ground but the RSI will not. The same argumentapplies in reverse, of course, when testing for market lows. The momentumlow, usually at the end of wave C, should be achieved with an RSI of 20 or less.


The market will then correct for the oversold condition and both prices and theRSI will rise. The subsequent price fall will take the market to a new low, but itwill not be confirmed by the RSI.

6. It is necessary here to be very aware of the relevant degree of pulse that is beingtraded. If a high-level up trend is in force, the trader should ignore short-termoverbought conditions and buy short-term oversold conditions; conversely fora high-level down trend.

7. The classic work is Wilder, J Wells (1976) New Concepts in Technical TradingSystems, Trend Research, Greensboro, South Carolina.

8. Elder, Alexander (1993) Trading for a Living, John Wiley, New York. In myopinion, this is one of the best books on the fundamental applications of tech-nical analysis on the market.

9. That is, +DI will probably give a non-confirmation at the peak of a bull move,and the –DI will give a non-confirmation at the bottom of a bear move.

10. Limited, that is, within the context of the hierarchical level being analysed. Thelast wave of a supercycle pulse may remain unconfirmed for years.

11. Rhea, Robert (1932) Dow Theory, Vail-Ballou, New York.


349

Part Four

The trader at work

350


351

24

The psychology of fear

INTRODUCTION

The foregoing chapters have developed two specific themes relating tofinancial markets:

� group psychology can be used to explain the existence of systemic,non-chaotic order in financial market prices;

� as a consequence, technical analysis is an entirely valid method ofanticipating future price action.

However, it is one thing to demonstrate the value of technical analysis;it is quite another to convert that information into effective, profitabletrading. The difficulty stems from two unavoidable aspects of humanpsychology. The first is that, no matter how careful an investor hasbeen in analysing a particular market, there is always the possibilitythat something has been missed. There is, therefore, uncertainty, anduncertainty means that a trader will inevitably experience some degreeof anxiety or at least be open to the possibility of experiencing anxiety.

The second problem is that investors tend to identify themselveswith their trading positions. That is, their sense of self-esteem andpersonal well-being become intimately bound up with whether or notthe positions make profits or losses. In the case of profits, the asso-ciated emotional pleasure is usually considered to be part of thebenefits of success; but it is a corollary of this attitude that losses auto-matically involve a great deal of emotional turmoil. There is thereforea danger that, in the face of any losses, natural low-level anxiety

caused by uncertainty will develop into fear. And fear (as demon-strated in Chapter 7) is likely to drive the trader straight into the armsof the crowd.

For most people, therefore, it is impossible to maintain an effectivetrading performance over long periods of time if high levels of fearand, in turn, stress are being experienced. Sooner or later, the physicaland psychological effects will either undermine the individual’shealth or it will begin to encourage serious trading errors. In bothcases, the only logical outcome will be a cessation of trading –whether temporarily or permanently.

A commitment to the use of objective entry and exit signals is anecessary condition for successful trading and investment insofar as itlimits the influence of the crowd. However, this commitment is not asufficient condition. The difficulty is that any system that allows atrader literally to see the recent move in prices automatically exposesthat trader to the essential catalyst of the crowd mentality – namelymoving prices themselves. It is therefore also necessary for marketparticipants to be able to deal with the psychological damage –whether actual or potential – that accompanies risk-taking. Thischapter therefore examines in more detail some of the appropriatemethods of limiting fear and of controlling stress while trading.

THE SUBCONSCIOUS MIND

Many people make the mistake of entering financial markets withoutany objective other than to ‘make lots of money’. While quitecommendable in itself, such an objective cannot activate the psycho-logical (and physical) mechanisms that contribute to success. Quitesimply, the objective of making money means nothing to that part ofthe mind – the subconscious – which is fundamental to the successfulcompletion of any task that we try to set ourselves.

Our knowledge of the processes that take place within the humanmind is, to say the least, incomplete. This is particularly true of thearea that is generally referred to as the subconscious – that is, the partof the mind that is beneath1 (or behind) the conscious mind. The diffi-culty is that, although we know the subconscious mind is there, wecannot usually access it directly during the course of our ordinarywaking lives. Its purpose is to make our regular behaviour as auto-matic as possible, so that our awareness is free to focus on new infor-mation.2 For example, once a child has learnt to walk, the processesare relegated to the subconscious so that it becomes increasingly

352 The trader at work

unnecessary to concentrate on the actual act of walking. The mind isthen freed-up to concentrate on other things. Insofar as it is possible tomeasure these things, scientists think that the subconscious accord-ingly accounts for as much as 97 per cent of our mental activity. Thisis a staggering figure.

The subconscious is, at its most basic, the storehouse for everythingthat we have ever experienced. However, it is far more than just amemory bank that can be accessed by the computer-like processingproperties of the so-called ‘logical’ part of our mind. It has its owndynamic, linked into the psychosomatic and autonomic nervoussystems3 of the human body, and it profoundly influences our liveswith its continuous stream of instructions and imperatives. It has itsown objectives – primarily linked to the survival of the physical andpsychological matrix of the individual4 – and unless it is properlyinstructed to accommodate new purposes, it will instead tend toprotect the integrity of its old learnt purposes against any new ones. Inshort, we are highly unlikely to be able to achieve our own personalobjectives unless we can enlist the help of our subconscious. We haveto teach the subconscious to accept change.

THE ROLE OF HABITS

The key to understanding the difficulties of initiating change is tounderstand the role of subconscious habits, particularly defensiveones. Habits are learnt mental, emotional and behavioural patternsthat are consistently reproduced in response to specific stimuli. It willbe remembered from Chapter 9 that there is a distinct three-phaseprocess to learning. The same process applies to habit formation: anindividual responds to a challenge by finding a coping strategy; if thestrategy ‘works’, it eventually becomes embedded in the subcon-scious; and it is then automatically invoked each time a relevantstimulus occurs. The famous example of Pavlov’s dogs is a good casein point. Generally, the purpose of a habit (as with any learntbehaviour) is to minimize energy and free-up the mind to deal withother matters, such as learning something new. However, there aretwo specific problems with habits: they may be inappropriate tocurrent circumstances and they interfere with the individual’s abilityto develop alternative strategies.5

This is particularly the case with habits that are carried over fromchildhood. For most people, external behaviour patterns are graduallymodified by social interaction, so that early habits are dispersed and

The psychology of fear 353

replaced. However, there is one area where modification is very, verydifficult. This is the area of emotional arousal in response to aperceived threat. The original threat may have been actual or illusory;nevertheless, it would have been treated as real. The brain and sympa-thetic nervous system then easily lay down pathways through the rela-tively untrained neuro-physiological network of the child. Thesepathways are subsequently the easiest to trigger. We shall say a littlemore about this in the next chapter, where we shall deal with the rangeof psychological responses to the threats implicit in market fluctua-tions. Here, however, it just needs to be registered that the first majorthreat faced by a child (which is almost always to do with their recog-nition of separateness from the mother) starts to lay down a pattern ofautomatic and inflexible responses that continue into adult life. Theprogramming therefore begins at an age that is beyond the scope ofnatural memory recall. Even those who can become aware of the auto-matic nature of their responses may not find their origin. However, thereal tragedy is that most of us are not even aware of these responses inthe first place; and, even if we are aware of them, we consider them tobe ‘natural’. Needless to say, these responses are more than likely tobe inappropriate.

THE RESPONSE TO A THREAT

As we saw in Chapter 2, whenever a threat is encountered, theamygdala starts to prepare the nervous system6 for ‘flight or fight’.Under some circumstances, the structure of the mind does appear toallow the forebrain some control over responses.7 However, thatcontrol is dangerously limited. Prior to some individual-specificboundary (largely determined in childhood), the control applies mainlyto external responses, but does not deal with the emotions that theamygdala triggers. Beyond the individual-specific boundary, the bodyreceives an injection of adrenalin and enters a full state of threat alert.

Obviously, a great deal can go wrong within this sequence ofevents. Even a low-level threat will trigger the essence of any habitualemotional response that an individual may have. A higher-level threatwill generate a strong emotional arousal, even if the associatedphysical behaviour is controlled. Unfortunately, failure to use theadrenalin by acting results in a build-up of unresolved tension (stress).Finally, a real shock will spontaneously trigger a full-blooded ‘fight orflight’ reaction, demanding of a full physical response. This latterresponse will be a very simple combination of body-oriented instinct


and reference-group regulated action.8,9 Whether it is fight or flightthat is the ‘path of least resistance’, the adrenalin will be used.

STRESS

If we translate these comments into the context of a financial market,it is likely that, at any one moment, large numbers of traders andinvestors will be experiencing either low-level anxiety because of thecontinuous threat of making a mistake or stress because a mistake hasbeen made but offsetting physical (that is, dealing) action has not yetoccurred. At any one moment in time, then, there is likely to be a greatdeal of unresolved tension. For each individual, this is likely to berepresented by one or more of a number of well-defined physical orpsychological symptoms, some of which are shown in Table 24.1.

These symptoms are also likely to be accompanied by negative (thatis, unpleasant) emotions of one sort or another. There are, of course, alarge number of such emotions that may be experienced, but the mostcommon ones that a trader is likely to experience and have to copewith are fear, anger, frustration, embarrassment and depression. Suchemotions use up valuable energy and divert attention away from thesolutions to the problem.

THE INFLUENCE OF EMOTIONS

The subject of emotions is actually a very complicated one and anypresentation will inevitably oversimplify matters. It is, however,


Table 24.1 Symptoms of stress

Physical Psychological

Shallow breathing IrritabilityRapid breathing Loss of self-esteemTightness in chest Inability to concentrateSweaty palms Loss of sexual desireClenched rear teeth Loss of memoryStomach problems Nameless fearsTense muscles Substance dependency

important that a trader has at least an elementary knowledge of theforces at work because, as we shall see, some degree of self-awarenessis an important part of the success programme. The main hurdle is thatour Western culture does not freely recognize the influence – or evenvalidity – of emotions, particularly in men. Indeed, the situation is sodire that huge numbers of people will, in all honesty, deny thatemotions – let alone anxieties – play any role in their decision-makingprocesses. As the analysis evolves, however, it is hoped that the under-lying realities will become a little clearer.

Emotions are motivating energies – a ‘giving out’, as it were. Theyare invariably aimed at an object. Further, they give meaning to ourexistence10 – life without emotions would be no life at all. Emotionsare therefore the primary determinants of the what, the how and thewhy of all activity. This is shown in more detail in Figure 24.1. First,of course, information from the environment is perceived by an indi-vidual. Second, it is automatically categorized and placed in a context.This context is the person’s beliefs about the structure of the world andabout their own function in it. Third, the process of categorizationarouses an emotional response. Fourth, the emotions focus energy ona limited number of learnt responses. Fifth, consciousness finallyenters the process and helps to determine the appropriate way ofcarrying out the learnt response.

Figure 24.1 emphasizes the point already made, that the process ofstimulus–response is largely handled by the subconscious. In otherwords, the broad strategy for dealing with a challenge is formedwithin the subconscious part of the mind. It is only after this point thatconsciousness enters the process to determine the details of anyresponse. The subconscious mind determines the strategy, theconscious mind determines the tactics.11

This means that the widespread conviction that behaviour isconsciously determined, and that emotions are therefore only the ‘fall-out’ from the process, is essentially false. This would probably notmatter very much if the process was just an interesting topic for debateamong academics. However, the process is actually about the waypeople understand the world; and if the understanding is at fault, thenit follows that the response is likely to be inappropriate.12 It is thissimple fact that explains so much: inappropriate responses result inunexpected outcomes; and unexpected outcomes, in turn, createnegative emotions and stress.


BELIEFS AND MEMORIES

As already described, the automatic release of mental energy, which istriggered by the initial stimulus, is filtered through a set of learntbeliefs. Beliefs give meaning to the stimulus. They are the accumulatedmemories acquired during the processes of socialization, academiclearning and personal experience. Furthermore, memories – almost bydefinition – have emotions attached to them. Memories are, in a sense,the ‘bricks’ from which our personalities are constructed and emotionsare the ‘cement’ that bind them together. Hence, the stimulus of ‘infor-mation in context’ will activate the emotions that are associated –singly, or in a blend – with the context itself. These emotions mayrange from being passive and almost imperceptible (such as feelings ofcomfort) to being active and totally explicit (such as a feeling of blindanger). Whatever their strength, however, the result is that – asemotions are psychosomatic phenomena involving the whole nervoussystem – a person literally becomes the emotion, or mix of emotions.

Once an emotional environment has been created in the psyche, itshould be apparent that the number of appropriate responses becomeslimited. An emotion will interact with the person’s learnt beliefs about


Beliefs:application of context

Beliefs:what is possible

Sensoryinput

Consciousthought

Behaviour

Emotions Neuro-physicalstate

Figure 24.1 The role of beliefs and emotions

the appropriate way to respond to the original information. Forexample, anger might find an outlet through either verbal or physicalchannels, depending on the outcome of previous experiences. The realproblem, however, is that the closer an emotion gets to one of theprimary ones relating to instincts,13 the narrower the range of behav-ioural options becomes. Hence, if the ‘fight’ emotion is triggered, the‘love’ emotion is automatically excluded.

These observations may now be applied to the simple example of atrader who is suddenly confronted by an unexpected change in prices(see Figure 24.2). The first thing that happens is that the unexpectedprice move is placed into the trader’s ‘habit of trading’ belief system.In other words, the trader recognizes the existence of a missed oppor-tunity or of an unexpected loss. There will immediately be a trace ofanxiety.14,15 If the price movement is a particularly large one, so thatthe implications are important, the trader is likely to experience fear –either the fear of missing further profits or the fear of makingincreased losses. The only possible reactions are either to resist themessage implicit in the price move (that is, that the adverse trend maycontinue) or to accept the implications and do something about it. Thisis a moderate version of the ‘fight’ or ‘flight’ response. If the former


Figure 24.2 The impact of emotions in trading

(that is, the ‘fight’) response is chosen, then there is obviously a riskthat the process is going to continue. In the face of a developing trend,the individual trader is likely to become wound up like a spring until,finally, the ‘flight’ route is chosen. Only then, when the emotion of fearhas finally forced the trader to deal, will consciousness be released todecide which stock will be traded and which broker will be used!

CONCLUSION

The purpose of this chapter has been to emphasize the roles of both thesubconscious and of emotions in all our activities. The influence of thesubconscious mind means that a great deal of our current behaviour ismechanical. In other words, much of what we do consists of automaticresponses to specific stimuli. The advantage of this is that currentchallenges can be dealt with using a combination of acquired skills,assimilated knowledge and sympathetic biological responses. Theproblem, however, is that an automatic response is triggered byaccessing an associated, emotionally charged memory of some kind.These emotional memories involve instinctual drives, racial, nationaland family histories, and any previous painful experiences. The emer-gence of a current emotion may therefore trigger reactions that are nolonger appropriate, even though they are based on similar pastemotions.

Inappropriate reactions are most likely to occur under conditionsthat generate anxiety and fear. Our basic reactions to anxiety haveusually been programmed into our psyches at a very young age, andmost of us are not even aware of them because they are outside of ourmemory recall. Anxiety is easily triggered once contact is made withtrading and investment positions in financial markets. We shall look atthis in more detail in the next chapter.

NOTES

1. The use of the word ‘beneath’ accords with accepted usage. However, it is morecorrect to understand the subconscious as being the context within which theconscious mind operates. Furthermore, in terms of motive power, the subcon-scious is far more powerful than the conscious mind. Research suggests that thesubconscious mind accounts for about 97 per cent of the mind’s total activity.

2. The conscious mind appears to be able to keep track of only seven items ofinformation at a time. This implies that the conscious mind is unable to copewith an infinitely complicated world by itself – hence the indispensability of


the subconscious. See Miller, George, The Magic Number Seven, Plus or MinusTwo, quoted in O’Connor, Joseph, and Seymour, John (1990) IntroducingNeuro-linguistic Programming, Mandala, London.

3. The psychosomatic nervous system deals with skeletal, muscular and sensoryactivity. The autonomic nervous system regulates the respiratory, circulatoryand other internal organ systems.

4. This incorporates the influences of the individual’s particular reference group.The subconscious will contain the belief system of the reference group and willtherefore be triggered by a threat to the whole group.

5. The power of habit is partly related to age. When we are young, we are veryopen to learning. As we become older, we become less open and more rigid:change becomes more difficult to accept.

6. That is, the ‘sympathetic’part of the autonomic nervous system, which is linkedinto the brain stem (see Chapter 2). The sympathetic system stimulates the bodyby means of the release of adrenalin. This part of the nervous system is to bedifferentiated from the ‘parasympathetic’ part of the system, which helps tocreate calmness by releasing relaxing endorphins.

7. Goleman, Daniel (1996) Emotional Intelligence, Bloomsbury, London.8. An individual threatened by muggers in an alleyway might respond by running

away. In a military battle, however, the same individual would likely fightalongside fellow combatants.

9. The point is that the neocortex does not necessarily step in either to control theresponses or to re-imagine the external world. Indeed, once primal passionshave been aroused, the neocortex may become an active participant in theprocess because its imaginings can validate and intensify the original emotionalresponse. Out of this specific weakness in the neocortex is born such obscenehorrors as human sacrifice and mass genocide.

10. Wilber, Ken (1996) A Brief History of Everything, Gill & Macmillan, Dublin,Eire.

11. Obviously, some emotions are more conducive than others to allow consciousthought to operate effectively: a passive sense of well-being, for example, ismore likely to accommodate effective communication between people than is,say, an active feeling of anger. The former allows for the wide use of language,reason and intuition, while the latter may tend only to allow a limited range ofverbal abuse!

12. This is precisely the criticism that can be levelled against modern economicmodels – particularly when they are used for economic policy targeting.

13. There is not complete agreement about the set of emotions that can be regardedas primal or instinctive. However, a general list includes fear, anger, lust or joy,sadness, acceptance, trust, disgust, expectancy, anticipation, astonishment andsurprise. These emotions correlate with eight fundamental adaptive behaviourpatterns that are directed towards survival and reproductive continuity. SeePlutchick, Robert (1980) Emotions: A psychoevolutionary synthesis, Harper &Row, New York.

14. Other things being equal, that is. The feelings being experienced when the actof perception takes place are of great importance. A person in an open,accepting state will tend to view a reduced range of information as a threat. Onthe other hand, a person in a closed, isolated state will tend to view a signifi-cantly wider range of events as being a threat.

15. There is, in all of us, a basic, deep-rooted belief in the unchanging nature of theworld. Our expectations are based on this belief. If outcomes do not match expec-tations, the implication is that something has changed. This causes anxiety.


361

25

The troubled trader1

INTRODUCTION

The previous chapter pointed the way towards a basic model ofpersonal investment behaviour that starts from a condition of anxietyand progressively winds itself up into high-level fear. Once anxietybegins to roll forward for any reason, it has the potential to createinappropriate actions. Inappropriate actions then result in unexpected,or unwanted, outcomes, which, in turn, create further anxiety.Eventually, the anxiety becomes fear. Persistent anxiety and fear, aswe have seen, result in individuals turning to a reference group – inthis case, the financial market crowd – for support. Immersion in thecrowd reduces an individual’s anxiety,2 but increases the likelihood ofmajor losses.

To be successful in trading and investment, an individual thereforeneeds to be able to ensure that the vicious circle of anxiety and inap-propriate actions does not get a hold. As previously mentioned, oneimportant part of the process is the use of objective entry and exitrules. The other part, however, involves direct management ofpersonal psychological states so that the integrative tendency is nottriggered. This is difficult to do, but any countervailing action is betterthan none.

The first step towards confronting the effects of anxiety was effec-tively covered in the preceding chapter. This is simply to recognizethat the danger exists. The second step, which we shall cover in thischapter, is to understand the specific way that anxiety influences eachof us personally. To achieve this understanding, we shall look at the

way that different psychological energies blend together to providebasic personality structures. The important point is that each person-ality ‘type’ is built around a very specific area of psychologicalvulnerability – a neurosis. And the third step, which we shall cover inthe next chapter, is to implement a strategy for neutralizing the effectsof anxiety before it becomes a threat.

THE THREE-PART MIND

In Chapter 2, we discussed Paul Maclean’s concept of a triune brain,consisting of the brain stem, the limbic system and the neocortex. Inessence, the brain stem deals with instinct, the limbic system dealswith emotions and the neocortex handles thought.3 The first point tomake here is that the human mind is dependent on all three parts of thetriune brain. The mind is a self-organizing system, just like any otherpart of nature, and it therefore simultaneously controls, coordinatesand consists of the three functions of instinct, emotions and thought.

The second, related point is that, according to analysts such asOscar Ichazo and Claudio Naranjo,4 the psychology of each individualis constructed from a blend of these three functions. However, there isstrong evidence that there are very subtle differences in the way thatwe each use them. Hence, some of us are biased towards instinctualactivity, others are regarded as being motivated by feelings oremotions and yet others are seen as being thought oriented. It is notalways possible to recognize these biases without careful analysis,but, whether they are obvious or not, they do help to determine theessential nature of the personality that we present to the world. Pleasenote that this does not imply any value judgements; it just means thateach of us has a bias towards one particular type of motivationalenergy and away from at least one of the others.

THE PSYCHOLOGICAL MATRIX

While each of us certainly tends to use one aspect of our three psycho-logical areas more than the other two parts, the evidence also suggeststhat there is an important subsidiary influence involving the other twoparts. This suggests that we can define basic psychological matrices interms of ranked combinations of the biases noted above. Hence, if wedenoted each of the separate motivations by a number – 1 for instinct,


2 for emotions and 3 for thinking – we should then be able to designatecharacter traits in terms of numerical combinations, such as 1–2–3,2–3–1 and so on.

However, the process of determining these numerical combinationsis somewhat complicated. First, some people are undoubtedly sodominated by their central motivation of instinct, feeling or thinkingthat the concept of a ‘combination’ becomes largely redundant. Inthese cases, only one descriptive number is relevant. Furthermore, itneeds to be emphasized here that, as we are dealing with mechanicalbehaviour, we are referring to the negative aspect of each of thecentral drives. Hence, instinct tends to be slothful, feelings tend not tobe authentic and based towards image needs, and thinking tends to bea refuge from fear.

Second, where motivations operate more clearly in combination,the ‘trailing’ motivation (the third number in each combination) isusually so weak that it is only relevant in so far as it identifies adefinite imbalance. In these cases, we may as well omit the thirdnumber from the combination, leaving just two numbers.

We are therefore going to denote each character ‘structure’either bya single number or by a double number. The first step, of course, is tocollate the relevant number combinations and, in order to do so, weshall use Figure 25.1. In this diagram, each of the three numbers isplaced at one of the corners of a triangle, starting with the number 1 atthe top.

BASIC PERSONALITY TYPES

Moving clockwise from the top of the triangle, and bearing in mindour earlier comments, we can make three combinations – 1–2, 2–3 and3–1. Moving counter-clockwise, we can make another three – 1–3,3–2 and 2–1. If we then include the solitary numbers, we have anotherthree – 1, 2 and 3. We have therefore arrived at a limited set of ninesingle paired numbers. These combinations can be arranged intogroups as follows:

� 1–3 for those who are instinctive, but are also influenced bythought;

� 1 for the instinctive;

� 1–2 for those who are instinctive, but are also influenced byemotions;

The troubled trader 363

� 2–1 for those who are emotional, but are also influenced byinstincts;

� 2 for those people who are emotional;

� 2–3 for those who are emotional, but are also influenced bythought;

� 3–2 for those who are thoughtful, but are also influenced byemotions;

� 3 for those who are thoughtful;

� 3–1 for those who are thoughtful, but are also influenced byinstincts.

PRIMARY BEHAVIOURAL CHARACTERISTICS

The second step is to isolate the primary behavioural characteristics ofeach of these combinations in order to understand what they actuallymean in practice. This has been done in Table 25.1. In the area ofinstinct, the bias involves attitudes towards relationships with theexternal world. These people are literally ‘gut-oriented’. In the area of


Figure 25.1 The psychological triad

emotions, the bias relates to feelings. These people are ‘heart-oriented’. Finally, in the area of thinking, the bias involves the attitudeto action in the world. These individuals are ‘head-oriented’. Thequestion now is what these combinations might be saying aboutsomeone. What does it mean to be a 1–2-type person or a 3–2-typeperson?

THE GUT-ORIENTED PERSONALITY

Let us look first at the people whose number combination starts with a 1.These are the instinctive, or ‘gut’-oriented, people.

The 1–3 people have an overdeveloped willingness to relate to theexternal world, but their feelings about themselves and their relation-ships are underdeveloped. They are likely to be self-confident andforceful, and they organize and lead. However, they may beaggressive, sometimes to the point of being destructive.

The 1 people tend to be out of touch with, or unaware of, the waythat they relate to their environment. Such people tend to be peaceableand reassuring. However, if they feel threatened, they may withdrawtheir attention from the threat in order to suppress feelings of anger.

The 1–2 people are going to have an underdeveloped ability torelate flexibly to the world. They are therefore likely to be principledand orderly, responding to the world’s requirements. However, theyare also likely to have very strong feelings about the need for orderand can, therefore, be perfectionists and punitive towards others.


Table 25.1 Primary behavioural characteristics

Type Character Bias

1–3 Instinctive–thoughtful Overdeveloped relating1 Instinctive Out of touch with relating1–2 Instinctive–emotional Underdeveloped relating2–1 Emotional–instinctive Overdeveloped feeling2 Emotional Out of touch with feeling2–3 Emotional–thoughtful Underdeveloped feeling3–2 Thoughtful–emotional Underdeveloped activity3 Thoughtful Out of touch with activity3–1 Thoughtful–instinctive Overdeveloped activity

THE HEART-ORIENTED PERSONALITY

Let us now move on to people whose number combination beginswith a 2. These are the ‘heart’-oriented people, who are defined by theway that their feelings are engaged.

The 2–1 people will have overdeveloped feelings, but theirinstinctual behaviour tends to be a limiting factor. They will be caringand generous (although potentially possessive and manipulative). Ifthey feel wronged, they may cut off a relationship altogether in orderto avoid negative feelings.

The 2 people are out of touch with feelings. Feelings tend to besuspended, while the immediate task is being concluded. These indi-viduals tend to be self-assured and competitive. Under pressure,however, they can be self-centred and hostile.

The 2–3 people will have underdeveloped feelings in the sense thatthey will not express them directly. Such people tend to express them-selves indirectly and therefore tend to be artistic, creative and intu-itive. They feel that they are special in some way. However, they canalso tend to be introverted and depressive.

THE HEAD-ORIENTED PERSONALITY

Finally, we come to those whose number combination begins with a 3.These are the thinkers, who are ‘head’-oriented and who maytherefore be defined in terms of their attitude towards activity.

The 3–2 people have an underdeveloped ability to do things. This isnot surprising given their neglect of the instinctive functions. Suchpeople tend to be perceptive and analytical in their approach to theworld, but they also tend to be somewhat eccentric and, when underpressure, paranoid.

The 3 people are out of touch with doing. That is, they acceptwithout question that activity is required of them. Individuals with thisstructure are dutiful and likeable. They are also dependent on othersand, under pressure, may be self-destructive.

The 3–1 people have a tendency to rush about because their abilityto do things is overdeveloped. They spend a lot of time planningfuture pleasures and are usually accomplished in the sense of having abroad experience of life. However, they can also be very impulsiveand somewhat excessive.


THE CHINK IN THE EGO’S ARMOUR

It is, of course, always dangerous to try and place precise labels onhuman nature. Indeed, the human psyche, when viewed correctly, isactually a process, not a structure. However, the process in each of ustends to get ‘stuck’, or becomes sluggish, around certain importantareas. At some stage – usually before the age of six – we confront ourseparateness in the world. As Oscar Ichazo puts it, ‘a contradictiondevelops between the inner feelings of the child and the outer socialreality to which he must conform’.5 The resulting trauma then sets upa centre of gravity within the psyche that, in turn, becomes the ‘deep-structure’centre around which our personality constellates. Accordingto Mr Ichazo’s work, there are only nine possible gravity centres andeach, quite simply, is the emotional memory of the trauma andpackage of responses that the young child used to deal with it.

The result is that the psyche will thereafter always be on the look-out for threats that are similar to the initial, archetypal one. And it willalways react with the same distinct package of learnt responseswhenever that archetypal threat is encountered. Importantly, however,not only does the psyche tend to react to all conditions of anxiety as ifthey were similar to the initial trauma, but the automatic nature of theresponses is not clearly recognized. If we are aware of them at all, wesee the responses as a strength, because they have served us so well inthe past.

In fact, however, the responses are also our central weakness. It isabsolutely essential to understand this in the context of financialmarkets. Our automatic responses ensure that we each have avulnerable spot within our psyches. Financial markets will seek outthat vulnerable spot with unerring accuracy and, accordingly, triggerautomatic and self-defeating behaviours. It is therefore worth whiletaking a little time to define these weak spots and the likely responsemechanisms in more detail. Then we will have taken a big steptowards doing something about them.

BASIC MOTIVATIONS

So, next, we can observe that each character structure has, as it were,two aspects:

� a defensive focus of attention;

� a likely mode of response to a perceived threat.


In other words, each character trait is like a coordinated radar andweapon response system. It surveys the environment for threat andthen responds with an automatic mechanical reaction if a threat isidentified. These two extra dimensions to the character traits areshown in Table 25.2.

The word ‘threat’ can, of course, mean a number of things. It can, forexample, mean a direct physical threat. Here, however, it means apsychological threat. Such a threat is experienced when there is adivergence between expectations (derived from subconscious beliefs)and actual outcomes. In financial markets, therefore, a threat isperceived if expectations of profit are not met. The resultant stress islikely to energize one of a number of possible automatic orcompulsive responses.

AVOIDANCE COMPULSIONS

Let us therefore move to the next stage by indicating the details ofthese compulsions. These are shown in Table 25.3. There is a differentcompulsion associated with each type of character structure. Eachcompulsion makes us avoid a particular condition or situation. Hence,for example, a 1–3 person will try to avoid being seen to be weak,while a 3–1 individual will seek to avoid physical or psychologicalpain.


Table 25.2 Character types and basic motivations

Type Character Focus of attention Response to threat

1–3 Instinctive–thoughtful Protection Aggression1 Instinctive of or1–2 Instinctive–emotional personal space withdrawal

2–1 Emotional–instinctive Protection Hostility2 Emotional of or2–3 Emotional–thoughtful ego deception

3–2 Thoughtful–emotional Minimization Anxiety3 Thoughtful of or3–1 Thoughtful–instinctive fear pretence

RESPONSE STRATEGIES

The question now arises as to what strategies each individual willadopt when faced with a situation that energizes their avoid-ance compulsion. That is, what happens when an individual feelsthreatened?

For each type of person, there is a specific initial defensive responsethey will make once a perceived threat is encountered. These areshown in Table 25.4. In very general terms, the instinct-orientedpeople will have to deal with the actual or potential emergence ofanger; the feeling-oriented people will have to deal with an impact ontheir self-esteem; and the thought-oriented people will have to copewith the actual or potential emergence of fear. Hence, for example, a1–3 individual will avoid being seen to be weak by fighting the sourceof the threat.


Table 25.3 Character types and basic avoidance compulsions

Type Character Avoids

1–3 Instinctive–thoughtful Being seen to be weak1 Instinctive Getting involved in conflict1–2 Instinctive–emotional Any form of imperfection

2–1 Emotional–instinctive Not being liked/recognized/loved2 Emotional Failing in chosen tasks2–3 Emotional–thoughtful Being just an ordinary person

3–2 Thoughtful–emotional Having insufficient data about life3 Thoughtful Making any wrong decisions3–1 Thoughtful–instinctive Physical or psychological pain

Table 25.4 Initial responses to threats

Type Avoids Initial response to threat

1–3 Being seen to be weak Fight the situation1 Getting involved in conflict Withdrawal1–2 Any form of imperfection Anger – usually suppressed

2–1 Not being liked/recognized/loved Dented pride2 Failing in chosen tasks Pretend that all’s OK2–3 Being just an ordinary person Envy the success of others

3–2 Having insufficient data about life Mental retreat3 Making any wrong decisions Fear3–1 Physical or psychological pain Make plans to retrieve situation

It is important to be clear that these are basic, or ‘core’, reactions.Each one will invariably be associated with other responses as part ofa reaction ‘matrix’. However, each core reaction constitutes a drivingforce for someone in any situation that is perceived to be a threat.Essentially, we all have a fundamental strategy to avoid the situationthat has caused us the most psychological pain in our early lives andevery time we are again confronted with one of these situations, wehave a deeply ingrained physical, emotional and mental strategy fordealing with the problem.

SOME AWKWARD PERSONAL QUESTIONS

Each person could – with a little introspection – determine their ownparticular character type from these lists. The reward would be someastounding insights into personal motivations. As a start, it may behelpful to look at the following list of questions. Each question, from1 to 9, corresponds to one of the personality types shown above. Thedifferent variations within each character type are far from complete,so the list only gives an indication of the responses involved.Nevertheless, the questions pinpoint some of the main characteristicsof the different types.

If you already participate in markets, simply ask yourself: when Iam confronted with a significant loss on an investment position, howdo I normally react? If you do not yet trade, ask yourself: when I amconfronted with a serious problem, how do I normally react? Youshould find that you can answer ‘yes’ to more than one of the ques-tions. However, one of the questions is likely to define very preciselythe reaction that matches you best. It will be the reaction that occursmost frequently, but, be warned, it may also be the one that you leastwish to admit applies to you.

1. Do you feel an aggressive need to organize something, such asyour desk, drawers or other people’s behaviour? Alternatively, doyou feel very sure that something else nasty is just about to happenbut you don’t know what it is, and so you don’t know what to do?(1–2)

2. Do you feel that the market has let you down and that you don’twant anything more to do with it? Alternatively, do you feel like afailure and tend to withdraw from people so that they won’t crit-icize you? (2–1)


3. Do you carry on as if nothing has happened, feeling that it’s goingto be OK next time? Alternatively, do you feel that things are notquite right, but pretend that things are great? (2)

4. Do you feel that you’ve made a mess of it yet again and, perhaps,want a drink to avoid getting depressed? Alternatively, do you startto fantasize about winning the lottery? (2–3)

5. Do you ignore the market and concentrate instead on carrying outa piece of research? Alternatively, do you arrange meetings andlunches with contacts or brokers? (3–2)

6. Do you feel cold inside and wish you were somewhere else?Alternatively, do you simply feel that things are going to get verymuch worse? (3)

7. Do you ignore the immediate problem and start looking for signalsor fundamental economic relationships that will ensure you get itright next time? Alternatively, do you think that you’re totallyinadequate for the job? (3–1)

8. Do you want to break something, throw the telephone down oreven hit someone? Alternatively, do you want to make someoneelse do something they don’t want to do? (1–3)

9. Do you feel amazingly relaxed when something goes wrong andwonder what others are making a fuss about? Alternatively, do youjust feel like looking after your own needs ‘for a change’? (1)

THE THREAT FROM FINANCIAL MARKETS

Now let us look at the situation for these nine character structureswithin the context of financial markets. Financial markets exhibitthree specific characteristics that are relevant to us in the currentcontext. These are that:

� they are moving continuously;

� there is an ever-present danger of losing money;

� they are subject to an infinite variety of influences.

The important point is that each individual is going to be particularlyvulnerable to one of these three characteristics of markets – that is,every individual’s defensive focus of attention (as indicated in


Table 25.2) is likely to be energized by one of these three character-istics. This will now be demonstrated with reference to Table 25.4.

FINANCIAL MARKETS AND PERSONAL SPACE

The first group of people consists of those who are primarily instinct-oriented. They will tend to have a very acute sense of self and protecttheir personal space. They will seek either to control or avoid anythingthat might intrude into this space – especially other people. In order tomaintain their security, the world is mentally classified into two parts:the part that is unchanging, and therefore unthreatening, and the partthat is active, and is therefore a potential threat.

Financial markets are moving all the time. They are full of unseenindividuals who will deal for reasons that cannot be discerned.Instinct-oriented individuals will therefore be confronted with thatwhich they fear the most – namely, an uncontrollable or intrusiveenvironment. They will react to losses with a desire to fight the market(1–3) or a withdrawal of attention and emotions (1) or suppressedanger (1–2).

None of these reactions is appropriate for successful trading. First,trying to fight the market when it goes against you boils down tothinking that you are right and the markets are wrong. This may wellbe true in the long run, but, as Lord Keynes said, in the long run, we’reall dead. Second, a withdrawal of attention in order to avoid angerimplies continuing to run bad positions, which could make things verymuch worse. Eventually, a failure to take account of what has beengoing on forces the individual to turn to others for help. Third, theprocess of suppressing anger means that important energy resourcesare going to be diverted away from coping with the market, stress isincreased and efficient trading is reduced. Ultimately, all that remainsis an overwhelming sense of inadequacy and guilt.

FINANCIAL MARKETS AND THE SELF-IMAGE

The second group of people consists of those who are driven by theirfeelings. These individuals are motivated by the need to protect theirself-image and will focus attention on gaining approval from others.Feelings of low self-esteem are generated when this approval is notforthcoming. Unfortunately, such feelings of low self-esteem often


derive from the fact that – in Western cultures anyway – we are trainedfrom childhood not to make mistakes. People who have responded themost actively to this sort of training are afraid to make mistakesbecause it makes them feel unattractive and the natural drive to findout about the world may even become suppressed.

Financial markets continuously present the danger, and the reality,of losing money. In fact, nobody can trade with a 100 per cent successrecord. Feeling-centred individuals will therefore invariably beconfronted with that which they desire least – namely, loss-makingtrades. The reaction is likely to be a sense of hurt pride (2–1), a falsepretence that a problem doesn’t exist (2) or envy of others who aremore successful (2–3).

Again, none of these reactions is conducive to successful trading infinancial markets. Attention is diverted away from coping with thereality of a loss-making trade towards coping with ego problems.First, hurt pride usually triggers hostility towards the market and mayinvolve a withdrawal from a relationship with the market. Second,ignoring the message of the loss and pretending that the position willcome right may eventually force the individual to turn to others foremotional support. Third, focusing attention on the success of others,which thereby highlights one’s own inadequacies, diverts attentionaway from solving the problem.

FINANCIAL MARKETS AND FEAR

The third group of people includes those who are characterized bythinking. Individuals in this general category are prone to experiencefear very acutely and are therefore motivated by the desire tominimize fear. Such people will focus their attention on informationand will be acutely aware of information that constitutes a threat. Thegreatest threat to security is that of making a loss and individuals whoseek to avoid losses are faced with the need to analyse an almostinfinite quantity of information.

Financial markets, therefore, confront thought-oriented individualswith precisely the situation they most want to avoid – namely, uncer-tainty. The result is that they will delay making decisions for as long aspossible (3–2), be frozen with fear and ignore threatening informationwhen a decision goes wrong (3) or avoid the issue, either by makingplans about future trading or even by doing other things (3–1).

Yet again, none of these reactions is appropriate for successfultrading. First, any delay in making a decision is likely to result in


either missed opportunities or higher-than-necessary losses.Eventually, the trader may be driven to frenetic action in a vain attemptto recoup the losses. Second, to be frozen by fear when a situation goeswrong means that the situation is likely to become very much worse.Ultimately, relief from fear will only be obtained by an aggressivecutting out of loss-making positions and, thereby, joining the crowd.Third, to pretend that the problem does not exist is merely to delay theday of reckoning. In the meantime, the trader may be forced into aphase of quite obsessive perfectionism in matters unrelated to a loss-making trade in order to avoid dealing with the problem.

THE EMERGENCE OF THE CROWD

So, to summarize, the critical problem is that financial markets arecontinuously generating all the most potent threats for every singleindividual all the time. As a result, the maximum number of people arefeeling threatened all the time.

Hence, despite promises of infinite wealth for the successful trader,the real experience is different. For the vast majority of people, theactual or potential emergence of losses is likely to generate negativeemotions and inappropriate defensive strategies.

It does not take a genius to recognize that the defensive strategiesdescribed are unlikely to rectify the situation. Put bluntly, as stressincreases, the attitude to dealing becomes less and less rational.Furthermore, personal psychological stability can become severelythreatened. Incorrect decisions are therefore likely to mount andpersonal life satisfaction is likely to deteriorate. Inevitably, the traderwill seek relief from the anxiety by identifying with other traders withsimilar views. In other words, a herd instinct is triggered. Ouravoidance compulsions therefore drive us into the welcoming arms ofthe crowd.

CONCLUSION

This chapter, in a sense, completes a circuit. In earlier chapters weargued that fear was the main catalyst for the formation of a crowd;and we showed how price fluctuations stimulated fear in financialmarkets and generated ordered price movements. Many people,however, would deny that fear (even fear of missing an opportunity)


ever forms part of their motivation to deal. Nevertheless, this chapterhas demonstrated that no one is actually immune: what happens is thatbasic anxiety is hidden from us by automatic avoidance compulsions.Indeed, most of these compulsions do not actually seem to be aresponse to fear at all. The truth, however, is otherwise.

As a result of the compulsions, an increasing number of peopleexperience a divergence between expectations and outcomes, soanxiety increases. Inevitably, the anxiety (or associated negativeemotion) becomes intolerable. At this stage, and whether we realize itor not, most of us end up as part of a crowd. Paradoxically, as part of acrowd, we can then all experience the extremes of naked fear togetheras the market eventually turns viciously against us. This is the natureof bubbles and crashes.

The question, then, of course, is how can individuals overcometheir avoidance compulsions and evade the magnetic pull of the crowdmind so that they can trade successfully in financial markets? As hascontinuously been emphasized, part of the solution is to adopt adecision-making process – a trading system – that generates buy andsell signals on an objective basis. In principle, the use of a systemgives the trader the potential to separate themselves from the herd.However, the use of a trading system is not a sufficient condition forsuccess. It is a necessary condition, but it only provides the workingtools. The difficulty is that any system that allows a trader literally tosee the recent move in prices automatically exposes that trader to theessential catalyst of the crowd mentality – namely moving pricesthemselves.

There are two basic possibilities for resolving this problem. Thefirst is to operate directly on the avoidance compulsions themselves.However, this can only be done through an intense and disturbingprocess of self-observation, which is undoubtedly a life-time processand not necessarily to everyone’s taste.

The second possibility is to focus attention instead on minimizingthe mismatch between expectations and outcomes (thereby reducinganxiety) and on managing negative emotions when they arise (therebycontrolling the integrative tendency). And so, it is to this latter optionthat we shall turn in the next chapter.

NOTES

1. This chapter is a significantly amended version of an article that first appearedin the winter 1995 edition of The IFTA Journal.


2. Psychologically, the individual returns to a state of ‘oneness’ with a significantother – here, the crowd. This reproduces, albeit in elementary form, the feelingstate of an infant in its early relationship with its mother.

3. Maclean, Paul (1979) ‘A triune concept of the brain and behaviour’, in Boag, T(ed.), The Hincks Memorial Lectures, University of Toronto Press, Toronto.

4. See, for example, Ichazo, Oscar (1982) Interviews with Oscar Ichazo, AricaInstitute Press, New York. Mr Ichazo is the source of a new perspective onhuman development, which places the human psyche in the wider context ofuniversal consciousness. Unfortunately, his work is still largely unpublished –mainly because he is not in favour of his work being used outside the context ofpersonal consciousness development. However, and for better or worse, theessential elements of his insights have been made more widely available viathe writings of Claudio Naranjo and Helen Palmer. The former willinglyadmits his debt to Mr Ichazo; the latter apparently does not. For those who areinterested in finding out more – particularly from a clinical perspective – I haveno hesitation in recommending Claudio Naranjo’s book (1994) Character andNeurosis, Gateways/IDHHB, Nevada City, California.

5. Oscar Ichazo, op. cit.


377

26

The psychology ofsuccess

INTRODUCTION

In this chapter, we shall address the basic personal requirements forsuccessful trading and investment. The analysis will be conductedfrom the point of view of someone who wishes to embark on a careerwhere personal account trading is essentially the sole source ofincome. If supplementary sources of income are available, then somuch the better.

The strong message – which cannot be stressed enough – is thatsuccess is directly related to our willingness to accept completeresponsibility for our activity in, and response to, markets. In this way,we can never be – and we can never regard ourselves as being –victims. Markets do not ‘happen’ to us, they do not ‘make’ us trade,they do not ‘make’ us miserable and stressed. Markets just are. It cantherefore be no one’s task other than our own to ensure that we arefully prepared to participate in this constantly moving and risky, butexciting and ultimately rewarding, environment.

BASIC REQUIREMENTS

In order to counteract the problems associated with anxiety and stress,it is necessary to adopt both an attitude and a lifestyle that are

conducive to regular participation in financial markets. This is partic-ularly true where trading is the sole (or main) source of income. First,a trader must be quite clear about their goals, or objectives, whenmaking a commitment to trade markets regularly. Second, the tradermust be able to sustain the energy necessary to maintain thatcommitment. Third, the trader must have a suitable method forachieving their pre-determined objectives (see Figure 26.1).

GOAL-SETTINGStress and negative emotions are essentially caused by a divergencebetween expectations concerning an outcome and the actual outcomeitself – either people don’t get what they want or they get what theydon’t expect. This is shown in Figure 26.2. Not only is the area repre-senting ‘achievements’ much smaller than the circle representing‘expectations’, but it also lies partly outside of that circle.

There is an old proverb that says if you aim at nothing, that’sexactly what you will hit. The implication is that it is vital to aim atsomething specific. There is, however, very little general under-standing of how this should be done.

The most effective way of establishing objectives is known as‘goal-setting’, or outcome-setting. Establishing goals has two clearresults:

� it significantly reduces the likely gap between what is expectedand what actually occurs and, therefore, contains stress in such away that it is manageable;


Goals

Energy Method

Figure 26.1 Trading prerequisites

� the technique actually increases an individual’s capabilities to asignificant extent by focusing energy on achievable results.

In Figure 26.3, these results are represented by the closer correlationbetween the circles representing ‘expectations’ and ‘achievements’ andby an increase in the size of the circle representing ‘achievements’.

GOALS FOR THE TRADER

Because of the intensely personal nature of goal-setting, it is not appro-priate to analyse the range of potential objectives relating to personalneeds, family needs and career needs in great detail here, but it is possibleto make some general points. The important starting point is to find out

The psychology of success 379

Achievements

Expectations

Figure 26.2 The cause of anxiety and stress

Achievements

Expe

ctationsGoals

Figure 26.3 The advantages of goal-setting

what it is that actually motivates you as an individual. Towards what doyour desires point?1 Quite simply, if you follow your deepest desires(when these have been properly elucidated), you will harness the mostpowerful energy available to you.2 The second step is to place thesedesires into a philosophical framework to ensure that there is no incom-patability between your needs and those of the community in which youlive and work.3 The third step is to make a plan of action, along the linesof the goal-setting procedures that will be outlined shortly.

If your family and social contexts are, indeed, amenable to a tradingcareer, then the relevant goals narrow down to deciding whether or notyour trading income is adequate and you are emotionally satisfied bythe processes of trading. In fact, the question of the adequacy of yourincome can be further subdivided into whether or not you are finan-cially secure and your trading methods are profitable (see Figure 26.4).

The three goals of successful trading, financial security andemotional satisfaction are actually contingent on one another: each isimpossible without the others. Thus, there will be no financial securitywithout successful trading; there will be no successful trading withoutemotional satisfaction; and there will be no emotional satisfactionwithout financial security.

PRACTICAL CONSIDERATIONS

It would be great if it was always going to be possible to start tradingwith a huge profit that then set you up for the rest of your career. Inpractice, however, the reality is likely to be a little bit more prosaic.


Successfultrading

Emotionalsatisfaction

Financialsecurity

Figure 26.4 The trader’s goals

The best advice that can be given is that it is absolutely essential topace yourself in the early period. Even rockets that are headed for thestars need to go through a process of acceleration before they can leavethe Earth’s orbit.

It will be very important to establish, from the outset, exactly whatlevel of income you are aiming for. This requires that you actuallyanalyse your outgoings in some detail so you know before you startwhat costs you need to cover. It goes almost without saying that youwill need to be practical about this. It may be necessary to start atrading career by amending your lifestyle (temporarily) in order to cutyour costs. Once you have done this, your income requirements arelikely to be more achievable.4

This, of course, raises the issue of what capital you have available tostart on a trading career. This is a very personal decision. As a rule ofthumb, for traders with no other source of income, it is probablynecessary to assume that you will be able to make at least 20 per cent perannum on your capital outlay. Hence, if your income requirements are£20,000 per annum, then you probably need to start with a capital outlayof £100,000. In good years, you are likely to earn considerably morethan this, but in a bad year you could earn less. The problem is that youmay start with a difficult year. For example, through no fault of yourown, trading conditions may just be very difficult, with low volatility.

The inherent riskiness of trading suggests that a measured approach isappropriate. There is, in fact, a strong case for starting to trade in smallamounts while actually pursuing another profession. In this way you canlearn without risking very much. Do you, for example, yet have theexperience to trade effectively? It may only take a year or two to learnhow to be able to go through the mechanics of trading, but it may take upto five years to learn whether or not you are able to cope with the tradingenvironment. Further, do you understand fully the technical issuesinvolved in the market(s) in which you intend to trade? The technicaland legal aspects of trading are becoming increasingly complicated andneed to be learnt and implemented carefully.

If you start with care, you will establish the base – in terms of bothfinance and experience – from which you can achieve serious results.

THE FIVE ASPECTS OFEFFECTIVE GOAL-SETTING

These practical considerations will need to be incorporated into yourgoal-setting procedures. They are designed to ensure that the subcon-


scious mind knows what to expect – that is, the subconscious isprogrammed with new, practical, beliefs about the future so that it canhelp rather than hinder you. For example, if you do not actuallybelieve that you can trade markets successfully, then you will simplynot be able to trade markets successfully.

The central issue is clearly to define the desired outcome itself.This outcome will be a combination of internal experiences andexternal behaviours that represent the desired state. For example, itmay be to become a successful trader. There are, in fact, five distinctconditions that must be met in order to establish a goal that will beeffective for you:

� It must be stated in the positive rather than in the negative. It isimportant to be precise about exactly what is required, rather thanmerely eliminate undesirable conditions. The subconsciousprobably already knows what you don’t want anyway, which iswhy it keeps projecting negative emotions at you.

� It must be ecological – in the sense of being consistent with yourrelationships with other people, with your environment and,indeed, with your beliefs about life. In other words, it must fitcompletely and naturally into your life and preserve balance.

� It must be stated in terms of a relevant context. Your subconsciouswill need to know exactly when, where and with whom you wantthe outcome. The realization of the goal must be placed within aneffective time frame. Obviously a goal that is to be realized withinthe space of a few months is going to generate more energy thanone that is to be realized within an ill-defined number of years.However, the time period must be practical. Doubling your incomein two years is a more practical objective than doubling it withinone month.

� It must be defined in terms of a sensory experience. It is importantthat you know what it will be like to have the outcome. Hence, youshould be able to visualize what it will be like to have the outcome.You should be able to see, hear and feel it in as much detail as youcan.

� It must be within your power to initiate and maintain the resourcesnecessary to produce the outcome. If you are unable to produce theresources or effect the necessary changes in the environment, thenthe project will not get off the ground. It would, for example, not bepractical to trade a market without sufficient capital.


CONVERTING DESIRES INTOACTUAL BELIEFS

These five conditions for defining an outcome are also the five stepsthat must be taken in order to programme the subconscious. If youcommit yourself to goal-setting, then it is necessary to go through thefive conditions outlined above, in sequence, one at a time. Further, ateach stage of the goal-setting procedure, the subconscious must befully engaged so that it will automatically accept the goal as being avalid call on its energy. In this way, desires become beliefs.

Much has been written about this mysterious process in recentyears.5 At one level, the very act of defining achievable out-turns in apositive way transmits important information to the subconsciousmemory. At another level, however, there is strong evidence tosuggest that certain techniques significantly enhance the process.These techniques are visualization, writing and affirmations.

VISUALIZATION

The subconscious does not communicate with the conscious mind vialanguage: it communicates instead in images and feelings. Hence, thesimple act of visualizing a goal as clearly as possible, with all its asso-ciated sensations, is treated by the subconscious as if it were reality.This is because imagination and memory actually share the sameneurological circuits. Visualization, in effect, creates a new realityfor you, so your nervous system and your emotions will behave as ifyour visualized situation already existed. Consequently, the visu-alized goal will quickly sink into your subconscious as a strongbelief. All that then remains is to transfer the belief from the innermind to outer reality.

To be really effective, it is recommended that the five steps outlinedabove are completed in one session (although the session can berepeated). You should take as much time as possible and be in asrelaxed a state as possible. Most authorities, in fact, recommend thatthe session be started by using a simple breathing exercise. Breathedeeply for a few minutes, all the while concentrating just on thedrawing in and flowing out of the breath. This has the benefit of bothdiverting the mind’s attention away from non-relevant issues andactively slowing down the brain’s basic rhythm. If the hoped-forfuture state is experienced in the mind’s eye in full detail – complete


with sights, sounds and smells – while in a relaxed state, it is morelikely to become the expected state than if you are tense.

Start by defining the desired outcome. It may be something like ‘Iwant to be a trader who earns £XX,000’. Keep it simple, so thatqueries and questions do not immediately arise. If all you have to startwith is a behaviour or state that you don’t want, then ask yourself whatit is that you would rather have. This avoids the spiral into asking‘Why?’ and blaming external circumstances.

Having stated the goal to yourself, check that you really do want it.Ask yourself specifically ‘Will I now accept this goal?’ If the answeris an unequivocal ‘Yes’, then you’re on your way, but watch for anydoubts that may emerge, particularly those that start ‘Yes, I want thegoal, but . . .’. If there are any doubts, go back to stating the goal in away that takes care of the doubt.

When you have a clear ‘yes’, state to yourself, in the clearestpossible way, the time when your goal will be achieved, the placewhere it will be achieved and with whom it will be achieved.

Then – and this is the crucial stage – visualize the goal in full detail.See it as being your reality.

Finally, confirm to yourself your own part in moving towards thegoal. See yourself gathering the necessary resources and completingthe necessary tasks. Again, see it as your reality.

It may actually be necessary to conduct this exercise more thanonce. Often it is helpful to have someone – such as a trained practi-tioner in neuro-linguistic programming – to help you through theexercise. The results undoubtedly warrant the effort.

WRITING DOWN AND AFFIRMATIONS

The process of visualization can be supplemented and enhanced bytwo powerful reinforcing techniques. The first is simply to writeyour goals down on a piece of paper once you have completed thefive-step process described above. The act of committing the goalto paper helps to ensure that the subconscious recognizes yoursense of purpose.

The second technique is designed to help the subconscious tocontinually orientate itself towards the goals that you have already setyourself. This technique involves the use of affirmations. Simply put,an affirmation is a statement, made to yourself, that affirms aparticular condition to be true. The affirmation is made regularly to


yourself, out loud if possible. Hence, for example, the simplestatement ‘Every day I trade according to my system trading rules’will be a powerful reinforcement to your goal of doing so.

There are, however, a few important rules to be followed whensetting affirmations for yourself. The first thing to remember is thatthey can be used only to manage beliefs, they cannot be used tomanage emotions (which are determined by beliefs). You willachieve very little by telling yourself that you’re feeling generouswhen in fact you are angry. You would need to focus on the belief thatis causing the anger.

The second thing is that the affirmation should be positive andpersonal. It should be couched in language that asserts a particularreality as definitely being true for you and does not implicitly leave anunanswered question. There is little value in either leaving yourselfout of the affirmation (‘life is great!’ – for whom?) or in just excludingundesirable conditions (‘my life is not bad!’ – then what is it?) Suchstatements have little or no effect on the subconscious. Third, thestatement should (if possible) be in the present tense (‘I am aneffective trader’) or, if it is in the future, include a time target (‘I shallhave an income of £20,000 per annum by my next birthday’).Otherwise, the mind will simply dismiss the affirmation as sheernonsense. The point is, you have to believe in it.

Goal-setting, combined with the act of committing the goal to paperand then backing it up with appropriate affirmations, is very simple.All that it requires is that you put aside a little time to implement theprocess. It is an understatement to say that your life could be trans-formed as a result. Those who have persisted with it have confirmedthat it is a very powerful technique.6

STRATEGY FOR ACHIEVING GOALS

Once a goal (or set of goals) has been set and started to become part ofa person’s belief system, it is important to concentrate attention on theprocess of achieving that goal. Goals belong to the future; the processis the immediate task in hand. Business plans are all well and good, butit is also necessary to marshal the necessary human, mechanical andfinancial resources into a cohesive strategy in order to achieve the plan.

The diversion of focused attention towards process is, in fact,critical. This is because there is an important difference between goalsand process that is often missed: goals are targets, not achievements;


and it is achievements that attract and manipulate emotional energy.If, therefore, attention is focused on what can be achieved now, hugeenergy resources can (and will) be released from unfulfilled desiresand from unpleasant emotions that, instead, can be used to deal withthe tasks of the moment. You will begin to live closer to the present.

STRATEGIES FOR TRADERS

The question for the trader in this context is whether or not it really ispossible to divorce the longer-term goal of profitable trading from thepotentially traumatic short-term effects of incurring losses. The answerlies in the making of two specific commitments. The first is thecommitment to use a technical trading system (the ‘method’) thatprovides automatic entry and exit criteria and incorporates moneymanagement principles. This will encourage you to remain disengagedfrom the emotional contagion of the marketplace and encourage you tolimit your risk. The second commitment is to the adoption of an attitudetowards yourself (the ‘energy maintenance’) that is supportive oftrading. The basic need is to maintain your energy levels in order to copewith the inevitable shocks and vicissitudes of the markets.

METHOD

The basic requirements for a successful trading technique are shownin Figure 26.5. They involve a technical trading system, the use ofstops and the spreading of risk.

In the first place, it is essential to use a trading system that generatesautomatic entry signals. These signals may be taken as suggestionsrather than as fixed instructions, but they reduce uncertainty by elimi-nating the need to search for further confirming information.

Second, it is essential to use ‘stops’ to exit a trade. Stops are levelsat which profits must be taken or losses realized once the marketbegins to move in the opposite direction to that required. Stops shouldbe regarded as instructions to deal and should not be cancelled oraltered. Their use will reduce stress by ensuring that, as far as ispossible, losses are not allowed to become too large to handle andprofits are not allowed to erode unnecessarily.

Finally, trading with specific amounts of capital, in more than onemarket, should help to spread risks so that the overall effect of the


volatility of profits and losses is reduced. In this way, ultimate successis not dependent on the profitability of the last trade. Consequently,your emotional fluctuations are also kept to a minimum.

A practical example of a profitable short-term trading methodology,which uses the three criteria outlined above, is set out in Chapter 27.

ENERGY

The final requirement is to maintain energy. By this is meant that thesuccessful trader needs to have the ability to maintain the energy andconcentration levels that are essential to pursuing profits in a fast-moving, high-risk operating environment (see Figure 26.6).

It is usually accepted as being logical that a successful tradingcareer requires clear thinking, physical energy and emotional freedom.The point that is often missed, however, is that physical, mental andemotional health are required constantly. Even highly successfultraders are likely to suffer from ‘burnout’ and thereby suffer from lossof income and/or ill health if they do not incorporate a regimen of goodhealth into their overall strategy.

PHYSICAL HEALTH

It seems to be a widespread belief (at least in the West) that good healthis something you either have or you don’t have. In other words, it’s a


Spreadrisk

Signals Stops

Figure 26.5 The method

question of luck. While there is inevitably an element of truth in thisview, it is far from being the whole story. Quite simply, if a suitableamount of care is taken, then the chances of becoming ill are signifi-cantly reduced. Unfortunately, it is on the question of overall healththat most traders suffer from myopia. Indeed, the trading culture oftenrequires that traders pay relatively little attention to the basic prin-ciples of personal healthcare. High-fat foodstuffs (often eaten atspeed) and alcohol, combined with long trading hours and a sedentarylifestyle, are the hallmarks of modern financial operators.

It is, of course, extraordinarily difficult to overcome the peer grouppressures that operate in financial markets, but, ultimately, it is aquestion of balance. If sufficient care is taken with diet, exercise andrecreation, then the body itself will try to remain healthy.7 It is, afterall, its natural condition.

MENTAL AND EMOTIONAL HEALTH

Physical health is, however, only a part of the solution – mental andemotional health are equally important. There are, of course, manyfacets to this particular subject, but the central feature of a balancedpersonality is the willingness to accept the truth about a given situ-ation. If there is a gap between beliefs about the world and the reality,then the result is likely to be inappropriate expectations and, hence, adivergence between expectations and outcomes. Such a divergenceresults in psychosomatic stress and negative emotions. In some


Effectivementally

Energeticphysically

Enabledemotionally

Figure 26.6 Energy

circumstances, a failure to recognize the truth of a situation is likely toyield obvious and dramatic results. As an example, if somebody isgoing to make a physical attack on you, it is important to your survivalthat you should recognize the threat straight away. Treating overthostility as an act of friendship is a sign either of madness or of saint-liness. The problem, however, is that the gap between beliefs andreality can be very subtle. If, for example, somebody makes a friendlyapproach to you, it is not always certain that you will recognize it assuch. You may assume that there is an ulterior motive. Indeed, it is alltoo easy to mistake a friendly gesture for a threat of some sort.

These examples are, of course, only two from among many possi-bilities, but it is obvious that a divergence between beliefs and realitycan be the cause of great discomfort. We have already argued that theprocess of goal-setting can significantly reduce the problem, but, ofcourse, this does not eliminate it entirely. It is still necessary to dealwith shorter-term (strategic) objectives, and these objectives may beexpected but not met. The result is likely to be stress.

STRESS

There are two points to be made here. First, the emergence of signs ofstress (including negative emotions) can be a very useful item ofinformation. They point unerringly to the presence of a weakness inthe strategy for achieving your goals. For example, a series of lossesthat eventually becomes too stressful will point to a problem with thetrading system.

The second point – one that is all too often missed – is that stressand negative emotions may be related to problems that have occurredin the distant past and, as such, may be inappropriate to the present.Each of our personal experiences – particularly as children – helps tocreate in us a set of beliefs that we subsequently use to guide our lives.We have already mentioned that beliefs – which give the worldmeaning for us – are trapped in a sort of emotional glue. Some of theseemotions are empowering in the sense of channelling energy intoproductive use. Others, however, are decidedly disempowering in thatthey divert energy away from the task of maintaining a high degree ofpersonal well-being. This is particularly true if the emotions are beingsuppressed in some way.

The trouble is that the negative emotions of our childhood are notonly suppressed during childhood, but they are also easily invoked


during adulthood.8 As children, most of us develop defensive habits todeal with our specific fears, but they are usually inappropriate to thechanging world in which we live. Hence, when a habit fails to work,very strong emotional reactions are generated. We discussed this pointin some detail in the last chapter and confirmed that our ‘natural’habits are, in fact, something of a hindrance when dealing in financialmarkets. We observed that the emotions that were most likely to beassociated with these habits were fear, hostility, anger and frustration.Modern psychoanalytical theory suggests that these primary habitsand their associated emotions may be reduced, or even eliminated, bybringing them out into consciousness.

It is probably fair to say that research into the dynamics ofconsciousness is only just beginning to reveal the healing powerinherent in the ability of the conscious mind to concentrate on parts ofthe subconscious. Goal-setting confirms that it is possible to manip-ulate the subconscious into accepting instructions from the consciousmind, rather than the other way around. This in itself is a massive stepforward. However, potentially more important is the possibility ofaltering the nature of our defensive habits by focusing attention ontheir causes. The focused conscious mind can act like a micro-surgicallaser beam in separating today’s reality from yesterday’s memoriesand, as a result, the former is not polluted by the influence of the latter.There are various ways of reducing the persistent demands of oldhabits,9,10 but the solution, quite simply, is a careful analysis of ourbeliefs about ourselves and relationships in, and with, the world.

RELAXATION TECHNIQUES

It was earlier observed that deep breathing can be used to establish astate that is particularly helpful for visualization techniques. Inessence, the brain’s rhythms harmonize and slow when we do this, andresearch has shown that it has a wonderfully relaxing effect on themetabolism of the whole body. Now, one of the symptoms of stress isshallow breathing. The breath is taken in high in the chest and thelungs are rarely fully emptied. Obviously this is appropriate to a shortperiod of fast breathing, when the body is oriented towards copingwith a threat. However, if shallow breathing persists for extendedperiods of time, the body tends to become de-oxygenated. This adds tothe overall stress levels.


A direct corollary of this is that slow and deep breathing can elim-inate many of the physical and mental symptoms of stress. Accordingly,a commitment to even a short period of slow and deep breathing eachday – preferably in the morning, before the day’s work has begun – canhave profound and far-reaching effects. A large number of books havebeen written on this subject and it is worthwhile investing in one ofthem and then implementing its recommendations.11

CONCLUSION

Genuine survivors in financial markets – that is, individuals who areable to make a living by trading successfully over sustained periods oftime – are likely to be the sorts of people who pay attention to theirphysical and psychological health. Unfortunately, however, themajority of people who trade markets will not do this until a stress-related crisis of some sort forces them to change.

In the meantime, it cannot be emphasized enough that, at the veryleast, genuine success in trading markets involves the adoption of atrading system. Without the discipline of such a system, the very bestefforts are likely to be doomed to failure. It is, therefore, to the subjectof trading systems we turn next.

NOTES

1. A useful starting point here is to ask yourself four questions:What would I do if I no longer had to worry about money?What would I do if I only had six months to live?What would I do if success were guaranteed?What have I achieved that has given me satisfaction and a feeling of impor-tance?In answering these questions, you will get a very clear indication of yourprimary motivations.

2. The late Joseph Campbell used to encapsulate this truth in the advice‘follow your bliss’. See, for example, Cousineau, Phil (1990) The Hero’sJourney: The world of Joseph Campbell, HarperCollins, New York.

3. Essentially, this means undertaking to participate in the society within whichyou live and work. See, for example, Peck, M Scott (1978) The Road LessTravelled, Simon & Schuster, New York.

4. Ideally, your income from guaranteed sources (such as from bonds) shouldcompletely cover your budgeted living expenses. Then you are independent.Your trading capital should then be a separate issue, so that a bad trading periodwill not interfere with your basic requirements.


5. See, for example, Bretto Milliner, Charlotte (1988) A Framework forExcellence, Grinder & Associates, Scotts Valley, California. This book shouldbe available from any bookshop that stocks a range of books on neuro-linguistic programming (NLP). See also O’Connor, Joseph, and Seymour, John(1990) Introducing Neuro-Linguistic Programming, Mandala, London.

6. See, for example, Alder, Harry (1994) NLP: The new art and science of gettingwhat you want, Piatkus, London.

7. There are many excellent books that address the problem of achieving balancein the area of health. See, for example, Chopra, Deepak (1990) Perfect Health,Bantam Books, London.

8. See, for example, Miller, Alice (1983) The Drama of Being a Child, Faber &Faber, London.

9. Peck, M Scott, op. cit.10. One of the clearest analyses of the beneficial effects of self-observation is

contained in Krishnamurti, J (1954) The First and Last Freedom, VictorGollancz, London.

11. An ideal and practical approach to stress is the excellent book by Loehr, James(1997) Stress for Success, Times Books, New York. For a more traditionalview, see Markham, Ursula (1989) Managing Stress, Element, Shaftesbury,Dorset.


393

27

The mechanics of success

INTRODUCTION

The use of a trading system is essential for success. The discipline thatit imposes simultaneously keeps a trader (relatively) immune from theemotional contagion of the marketplace and enables decisions to bemade on the basis of objective criteria. There are, of course, almostinfinite numbers of systems that can be used, depending on yourpersonal preferences and the technology that is available to do thetask. The falling prices of very powerful personal computers and theready availability of automatic datafeeds make it possible for an indi-vidual to compete in sophistication with some of the biggest financialhouses. It is not, however, essential to use such technology. It can bejust as rewarding to track markets using nothing more than graphpaper – indeed, it can be validly argued that such an approach wouldactually give you a better ‘feel’ for the market. Be that as it may, thereare three basic criteria for effectiveness that need to be spelt out:

� a system needs to be profitable;

� it needs to be proven;

� it needs to be personalized.

SYSTEM EFFECTIVENESSThese three criteria are not mutually exclusive. The point of drawingattention to them is that you – as the trader – have to be satisfied that

the trading system being used actually meets your requirements. Asystem that leaves areas of significant doubt will not serve the goals offinancial security and emotional well-being. It is also important to becomfortable with the operations of the system.

It goes almost without saying that a system has to be able togenerate profits, but there are two related issues that need to beaddressed. The first is whether the system is going to be used tosupplement other sources of income or is to be the sole source ofincome. In the former case, the resources devoted to the initialresearch need not be burdensome; in the latter case, however, a signif-icant amount of time and (often) money does need to be spent inchoosing a suitable system.

The second issue that needs to be considered is whether you want toundertake a highly active trading profile in a limited number ofmarkets or deal less frequently, but in a large number of markets.

Once these objectives are clear in your mind, then the effort can bemade to establish a profitable system. In the initial stages, prof-itability obviously has to be tested on the basis of historical data.However, obviously, the real test of a system is its actual performanceunder live conditions. Any system will inevitably be buffeted bytrends that are longer-term than those for which the system wasdesigned. It is at such times, after a period of underperformance, thatsubjective (and invariably fatal) panic overrides can be forced intoplay. It is important, therefore, to test out the system in real time –but, initially, without making a massive financial commitment. Thiswill help to ensure that the system itself can cope with the actualspeed and volatility of the markets being traded, and that you cancope mentally and emotionally with those periods of time when thesystem generates losses.

This problem of coping is an important one and should not beunderestimated. In Chapter 24, it was argued that the emotional strainof dealing with losses might, in the final analysis, simply be too much.It is for this reason that care should be taken to ensure that the systemis the ‘right’ one in terms of feel and familiarity. Obviously, this isvery subjective, but then so it should be because it is a question ofconfidence and commitment. For example, most traders are morelikely to trust a system that they have designed themselves than onethat is essentially a ‘black box’ signal generator. This is not to say thatthe latter are ineffective – far from it. It is just that such systemsinevitably raise strong doubts when they start to make losses, becausethe cause of the losses may not be apparent straight away.


SYSTEM DESIGN

An effective system – that is profitable, proven and personalized –needs to have three elements:

� it should generate potential entry signals;

� it should define specific exit points;

� it should allow the spreading of risk.

There are obviously quite a few possibilities here, so, in order toexpedite the task of explaining the system requirements, we shall usea simple example. The example is a short-term trading system that canbe used in almost any liquid market. It is an effective system, but is notperfect. Its value here is only to demonstrate some of the ideasinvolved. All that is required is access to a price series, where eachprice is registered regularly at an hourly interval. If it is possible torecord the spread of prices during the course of an hour – andtherefore record the highest price, lowest price and last (‘closing’)price for each hour – so much the better. It is assumed that theseprices, and the associated indicators, can be plotted in graphical form.

The trading signals for this system are generated at the end of eachhour’s price movement (the closing price is the important one). If abar chart recording hourly spreads is used, then it is possible to checkfor warnings of an imminent signal by observing the behaviour of thecurrent price and comparing it with potential end-period signal levels.

THE SYSTEM STRUCTUREThe first task is to calculate two arithmetical moving averages –namely, a 26-hour moving average of the hourly closing prices and thehourly equivalent of a 21-day moving average of such prices (that is,21 times the trading hours in a day). The shorter average highlights theshort-term trend and the longer one relates to the longer-term, moreslow-moving, trend. Both averages are used to determine whether thecurrent price is bullish or bearish within a particular time frame. Acurrent price above a moving average will be bullish; a price belowthe moving average will be bearish.

Although it is unnecessary in a simple system such as this, themoving averages can be supplemented by an ‘average directionalindicator’ (ADX). This was discussed in Chapter 23. The ADX givesan indication as to whether or not the market is actually trending.

The mechanics of success 395

The second task is to calculate a momentum indicator that will giveyou both an indication of reversals in momentum and a tool torecognize non-confirmations. The most popular momentum indicatorin modern markets is the ‘moving average convergence–divergence(MACD) line’, which we also discussed in Chapter 23.

This indicator measures the difference between two movingaverages. When plotted as a graph, the line oscillates slowly up anddown as the two averages first diverge from each other, then convergeon each other and finally cross over one another. The original versionuses exponential moving averages, so that more weight can be givento recent time periods than to older ones. However, quite acceptableresults can be obtained by using simple arithmetical averages. Therecommended time periods for the two moving averages are 13-hourand 34-hour.1

When the MACD indicator is rising, the implications are positivefor prices: if the indicator is less than zero (that is, the 13-hour averageis still below the 34-hour average), then the market is only potentiallybullish; if the indicator is greater than zero (that is, the 13-houraverage is above the 34-hour average), then the market is actuallybullish. The converse holds true if the MACD indicator is falling: ifthe indicator is above zero, then the market is potentially bearish; ifthe indicator is less than zero, then the market is actually bearish.

The third task is to calculate an indicator known as the ‘signal line’.This indicator is a moving average of the MACD line. Again, the besttype of average for this purpose is an exponential average, butacceptable results can be obtained by using an ordinary arithmeticalaverage. The recommended period for this average is eight hours.2 Inprinciple, a definite recommendation to enter a trade is generatedwhenever the MACD line crosses the signal line.

Once the mechanics for calculating these averages have been estab-lished, all that is necessary is to set out a basic set of entry and exitrules, then use them. These rules are obviously variable, but oneparticularly useful set of rules is given below.3

THE ENTRY RULE

A position is opened when the MACD line breaks the signal line andone of the following supplementary (or confirming) signals occurs:

� momentum has not confirmed the recent price low (in the case of abuy signal) or high (in the case of a sell signal);


� the last price crosses the 26-hour moving average in the samedirection as the momentum signal;

� the last price signals a break-out from a valid price pattern;

� the price is reversing away from a golden ratio objective.

THE EXIT RULES

Half of the trading position should be closed when the price hits atarget calculated from golden ratio expansions or contractions, and ifone of the following criteria is also met:

� the golden ratio target is also at a support/resistance level;

� momentum is diverging from the movement in absolute prices;

� momentum is at an extreme and the ADX is not rising.

All (remaining) open positions should then be closed if either a pre-determined stop loss is triggered or if the MACD line breaks the signalline in the opposite direction to the trade.

CLOSING AND REVERSING POSITIONS

If the current price is above the (hourly equivalent of the) 21-daymoving average, then closing a short position automatically impliesopening a long position. Thus, if the current price is below the 21-daymoving average, then closing a long position implies opening a shortposition. Otherwise, the normal entry/exit rules apply.

It is also worth emphasizing that reversals away from the goldenratio objectives can be used to anticipate this effect.

INVESTMENT FUNDS

Within the context of this short-term trading system, it is advisable touse a ‘standard’ amount of funds for each trade. This standard unitshould be used every time a signal is given. However, there are twopractical alternatives. The first is to allocate half of the standardamount whenever the MACD entry signal (the primary momentum


signal) is given. The rest of the funds can then be allocated when one ofthe supplementary signals is generated. The second alternative is to waitand allocate all the funds when a full set of signals (a primary signal,plus a supplementary signal) is produced. This alternative might, in anycase, be usefully adopted when markets are particularly volatile.

If an average directional indicator is being used, some of the direc-tional uncertainty is removed. Accordingly, full amounts can be allocatedif the ADX line is rising, while half units should be used if it is falling.

THE THREE-WAVE MARKET MAPThe entry and exit signals can be used without reference to longer-termtrends. However, the knowledge that all price patterns are based on theprice pulse introduces tremendous scope for adding a valid subjectivedimension to trading decisions.

The archetypal nature of the price pulse means that a clear three-up/three-down base or top pattern is likely to emerge just prior to ahigher-order impulse wave.4 The pattern can be used to construct amarket map of the future, with appropriate objectives calculated fromgolden ratio expansions. Quite obviously, the larger this pattern, themore powerful the subsequent impulse wave is likely to be. The mapwill help determine whether or not longer-term investment positionsought to be opened, and, if so, what amount of money ought to be allo-cated to the trade.

If the emergent pattern is considered to be particularly strong – thatis, if the pattern has taken some time to develop and there are groundsfor supposing that the fundamental background picture is changing –then serious consideration should be given to committing somemultiple of the basic investment unit to the trade. This multiple shouldthen be invested when a short-term trading signal is generated. Ifrequired, the amount may be split into two equal units, so that half isinvested on the primary signal and the rest allocated to the supple-mentary signal.

It follows that shorter-term exit signals should not trigger a fullclosing off of all positions. This should only occur when the calculatedobjective is hit.

SPREADING RISKPart of the mechanics of success involves spreading risk, so that a lossdoes not incapacitate your trading programme. There are two basic


ways of doing this. The first is to trade in either a number of marketsor in a number of different stocks in the same market. The point is thatthe variety should be sufficient to ensure that any losses are going tobe more than offset by profits elsewhere, but the variety should not beso great that you cannot cope with all the information and activity.

The second way of spreading risk is to distribute it over time, ratherthan across markets or stocks. In other words, it may be moreacceptable to wait for those trades where you expect the outcome to beparticularly dynamic (because of a significant base or top pattern).The idea would be to keep a close watch on a number of differentmarkets, but only trade when conditions are right.

MONEY MANAGEMENT

Whichever approach is suitable for you, there are two cardinal moneymanagement rules that need to be obeyed5. The first is that no tradeshould allow you to lose more than 10 per cent of your capital. If youare trading equities, bonds or currencies, a ‘stop loss’ (see below)should be triggered well before the 10 per cent limit is hit. Suitableadjustments should be made for leveraged futures positions. Inoptions markets, no more than 10 per cent of your capital should beplaced on a single trade because there is a chance that the option mayexpire worthless. The 10 per cent rule should also be applied to‘recovery’ situations – where the company has been facing severefinancial difficulties, but is now considered to be over the worst.

The second money management rule is that no more than 20 percent of your capital should be placed in a single investment. This willensure that, if you are spreading your risk across markets or stocks,you should be able to hold at least five positions.

If these rules are adhered to, then it will be possible to weatherstormy times and still be able to take maximum advantage of prof-itable trading opportunities when they present themselves.

OPENING POSITIONS

We shall now discuss the signals themselves in a little more detail.The entry rules essentially rely on a reversal in momentum (theMACD line) to establish the basic signal. However, it is still possiblethat the reversal will be too short-term to provide a profitable trading


opportunity – hence the use of supplementary or confirming signals.The first signal relies on the concept of ‘non-confirmation’ that weanalysed in Chapter 23. If the primary momentum signal is beinggiven on a reversal from a price extreme, and there has been a non-confirming divergence between the price extreme and the MACDline, then the signal has a greater chance of being valid.

The second supplementary signal relies on the argument that themoving average approximates the trend in the market, and that a pene-tration of the moving average therefore indicates a change in trend.However, the signal has a greater chance of being valid if the movingaverage itself is in the process of reversing.6 The use of a movingaverage in this way is particularly useful when the primarymomentum signal has been generated under very choppy tradingconditions.

The third signal allows for the break out from congestion areas,holding patterns, recognizable reversal patterns and trend lines. Thesepatterns are the basis of much traditional technical analysis. In a sense,they are late signals because prices will already have moved someway before the signal is given. Nevertheless, such signals mayconstitute the last piece of information in the movement away from anarea of very volatile price action.

The fourth signal covers those cases where the price has reached avalid price objective and is reversing away from it. First, the pricemay have achieved a retracement objective of 38.2 per cent or 61.8per cent of the previous move, and may be moving back into the maintrend. An MACD entry signal should catch such a renewed trend.Indeed, it is particularly powerful for the 61.8 per cent retracement,where the market has re-tested the previous high or low. Second, theprice may have reached a golden ratio objective for an impulse moveand may be reversing into a new trend. The MACD entry signalshould allow you to catch the beginnings of this new trend, too.

CLOSING POSITIONS

It is all very well entering trades successfully, but it is obviously alsoimportant to ensure that subsequent profits are not allowed to erodeunnecessarily and that any losses are kept to an absolute minimum. Tothis end, the need to follow strict rules about closing positions cannotbe emphasized enough.


As discussed earlier, there are psychological influences thatencourage traders to take profits too early and realize losses too late.There is also the related problem of identifying with trading positionsso that self-esteem becomes dependent on the outcome of the trade.Both of these sets of problems can be reduced, or even eliminated, byadopting rigid exit criteria.

STOPS

The cardinal rule is that every trade must be accompanied by aprecisely defined set of exit rules. In other words, when a trade isentered, a clear decision has to be made about the conditions underwhich the trade must be closed. There are actually two situations to beaddressed. The first of these concerns the emergence of a loss on atrade. The question to be answered is, when is the trade definitelywrong so that it is necessary to stop the loss? The answer dependspartly on the circumstances under which the trade was opened. Hasthe original reason for entering the trade been aborted? Was the entrysignal so temporary as to be false? However, the answer also dependspartly on what level of losses can be absorbed for any one trade. Thisis a particularly important consideration in volatile markets, such asforeign exchange.

As already indicated, at no stage should more than 10 per cent ofyour capital be at risk on any one trade. A loss in excess of this amountwill seriously damage your ability to continue trading. The level atwhich a stop loss must be triggered should be determined at the timethat the trade is opened. Once a stop loss level has been determined, itshould neither be cancelled nor altered.

MECHANICAL EXIT SIGNALS

The second situation to be considered is the more fortunate onerelating to the conditions under which profits must be taken.Mechanical exit rules will ensure that profits are regularly ‘booked’and will also encourage you not to feel frustrated or annoyed if themaximum profit is missed. You never go broke taking profits,however small those profits may be. Having taken a profit, you canthen move on to the next trading opportunity without undue concernabout how successful (or unsuccessful) you’ve been.


CALCULATED TARGETS

One of the powerful messages from this book is that mechanical exitrules are substantially enhanced by incorporating calculated targets.This is because such targets help to avoid the premature closure ofwinning trades and can help generate an exit signal close to the pointof maximum profit. The only targets considered here to be valid arethose calculated from the golden ratio. Nevertheless, they should notbe used by themselves, simply because there are always various possi-bilities. Consequently, only half the position should be closed when avalid target is reached and then only if one of three other supportiveconditions is met – that is, either the calculated target must previouslyhave been associated with a price reversal or there is a potential diver-gence between prices and price momentum or the market is very over-bought or oversold.

HIGHER-LEVEL TRENDS

When calculating the potential targets, there is one very importantconsideration that must be taken into account. This is that the energyfrom the current level of the price pulse is derived from, or qualifiedby, the energy of higher-level pulses (see Chapter 5). Golden ratioobjectives for the short-term trend will usually work, but may bequickly washed away by the power of the superior trend. This impliesthat two factors need to be taken into account. First, the power ofhigher-level trends may cause prices to undershoot or overshootobjectives. Usually, the objectives are very precise, but a leeway of+/– 1 per cent should be allowed. Second, the influence of higher-level trends means that you will need to be aware of the longer-termcontext within which you are operating. Reference to recent importantturning points in the 21-day moving average of prices (see below) willusually provide a very good indication of this context.

However, when there is some uncertainty – for example, when it isnot immediately clear whether or not the market is just beginning acorrection to the most recent phase of a trend or is just about to begina major reversal – the correct approach is to be cautious, follow thegolden ratio signal on half of the position and be ready to re-enter themarket when a valid entry signal is generated. It may even be fair tosay that the real power of the application of calculated targetsprobably only comes with experience.


FINAL EXIT

Exit rules based on the golden ratio objectives will be very effectivefor a large number of – or even on most – occasions. However, oncethe market begins to reverse away from a target, then (whether or notany of the supportive conditions have been met) open positions shouldbe closed when the MACD line cuts the signal line. No other factorsneed to be taken into account.

CLOSING OLD POSITIONS ANDOPENING NEW ONES

There are two possible problems with these exit rules. First, if therehas been some uncertainty over the level of the price pulse beingtraded, there may be occasions when positions are, in effect, closedprematurely. If the dealing rules are followed consistently, then thisshould not actually matter too much as a re-entry signal will inevitablybe generated when the underlying trend re-asserts itself.

Second, there are occasions when the short-term trading systemdoes not immediately show whether or not the closing of an oldposition should be accompanied by the opening of a new one in theopposite direction. In order to circumvent this uncertainty, it is a goodidea to use a much longer moving average to define the backgroundtrend of the market. Without doubt, one of the best averages for thispurpose is a simple arithmetical 21-day moving average. If hourlycharts are being used, then it is a simple matter to convert the 21-dayaverage into the corresponding hourly average. The market trend isthen defined by the current price in relation to the moving average:

� if the price is above the average, and the average is rising, then thetrend is bullish;

� if the price is below the average, and the average is falling, then thetrend is bearish.

The rule for opening and closing positions is then quite simple:

� if the market trend is bullish, then closing shorts implies openingnew longs;

� if the trend is bearish, then closing longs implies opening newshorts.


As indicated above, using moving averages in this way may be supple-mented by using an average directional indicator. If the ADX is rising,then the market trend (whether up or down) is intact. Under thesecircumstances, closing contra-trend trades should be accompanied bythe opening of pro-trend trades. If the ADX is falling, then the marketis not trending and a more circumspect approach is required.

CONCLUSION

The basic system outlined in this chapter is far from being the lastword in profitable trading systems. The aim here has been to indicatehow a profitable trading system can be developed. For the individualtrader, great insights can come from just trying to develop a person-alized trading system. Furthermore, backtracking through all thehistorical signals and identifying their validity encourages great confi-dence in the system.

However, it should be clear from the analysis that profitability islikely to be enhanced by a genuine understanding of how marketsbehave, and by use of the golden ratio to determine price objectives.These, in turn, encourage patience while natural forces evolve andencourage confidence that expectations will be fulfilled. These are twoof the primary advantages of goal-setting, which we outlined inChapter 26.

NOTES1. Different analysts use different time periods for the calculation of the expo-

nential moving averages. The original ones recommended by Gerald Appelwere 12-period and 26-period averages respectively. The author’s ownresearches suggest that 13-period and 34-period averages may give betterresults.

2. Many analysts use nine-period averages.3. This particular trading system has been used by the author and is one that has

been refined over a long period of time.4. See Chapter 12.5. These rules are for general guidance only. In modern, highly leveraged markets,

great care needs to be taken with risk control. One of the best books on thesubject is Elder, Dr Alexander (2002) Come Into My Trading Room, JohnWiley, New York.

6. The moving average signal is unlikely to be correct if it is generated when themoving average is still moving sharply against the direction indicated by thesignal.


405

28

Summary andconclusions

An analysis of universal laws can provide us with an explanation ofstock market dynamics. The need individuals have to obey universallaws and participate in greater wholes is the key to a fuller under-standing of all social, economic and political activity. In the currentcontext, it allows us to ‘explain’ the stock market phenomenon byforcing us to adopt the assumption of non-rational crowd behaviourby market participants.

The important point to remember is that the ownership ofinvestment positions in the pursuit of stock market profits creates a‘crowd’ mentality. The decision to make an investment (or disin-vestment) may have been arrived at rationally, but, once a financialexposure exists, so does the potential for stress and fear. Pricebehaviour becomes the arbiter of investment decisions and pricemovements can therefore trigger responses from investors that have alarge non-rational dimension. This non-rational response is thebinding force in the crowd situation.

Once a crowd has been created, it has a life cycle of its own. In thecontext of a financial market, the crowd’s behaviour during this lifecycle can be measured by changes in prices and sentiment. The formercan be tracked using price–time charts, while the latter can be judgedby plotting variables such as volume, open interest or momentum.

While a crowd exists, it has natural metabolic rhythms that haveboth an internal and an external dimension. The internal rhythms are,by definition, specific to the particular crowd and are reflected inregular, cyclical fluctuations that can be found in the price–time

charts. At the same time, however, the crowd has to oscillate in a limitcycle relationship with its social and economic environment. Some ofthe environmental oscillations are themselves ‘regular’, and so thecrowd oscillates in harmony with these external rhythms because,essentially, it has ‘learnt’ to do so. In both these situations, where thefluctuations are regular and continuous, it is possible to use observedcycles to forecast the timing of turning points in financial marketprices.

Most of the environmental oscillations, however, will appear as‘shocks’ to the crowd. They may consist of either pro-trend infor-mation or contra-trend information – pro-trend informationencourages existing trends; contra-trend information generates acorrection; both generate a three-stage fluctuation as the crowd seeksto assimilate the information.

This three-wave response is the basis of what we have called the ‘pricepulse’. Each phase of the price pulse is defined by the golden ratio and itsderivatives. The golden ratio has long been recognized as an intrinsicpart of nature in terms of both the relationship between different aspectsof the same spatial structure and the growth and learning processes ofthat structure. It follows that price movements in financial markets aresubject to the ‘constraints’ of the golden ratio and it is therefore possibleto forecast the extent of price movements using this ratio.

Finally, the alternating up and down movements created by thecombination of rhythmic cycles and contra-trend shocks are reflectedin a limited number of price patterns. These have long been identifiedby traditional technical analysis and can be used to deduce tradingsignals that have a low risk of being incorrect and a high probability ofbeing profitable.

Of course, no system is perfect. There will be occasions when themessages being given are either inconclusive or conflicting. On theseoccasions, investors should not anticipate reversals if to do so means‘moving away from base’. At the simplest level, this implies that long-term investors should always stay in the market until definite sellsignals are given, and that short-term investors should always stay outof the market until definite buy signals are generated. On most occa-sions, however, technical analysis will confirm the presence of areversal while it is in the process of developing.

The crucial requisite for wealth accumulation is for a trader to beproperly positioned for large market movements. It is important toremember that large movements are created, essentially, by the failureof the majority of investors to anticipate all developments fully, andby the consequent reaction en masse to largely unforeseen events.


Substantial price movements simply cannot occur if everybody isgetting the market right all of the time. Technical analysis focusesattention on what the majority of investors are doing at any givenmoment, while simultaneously enabling successful investors to standaside from the crowd mentality. It is a rational approach to a non-rational environment.

Hence, technical analysis should, at the very least, ensure that nomajor mistakes are made. When used properly, the techniques will notgive strong buy signals just prior to a price collapse, nor strong sellsignals just prior to a rally. More generally, there are always clearwarning signals that a terminal juncture is approaching and thesesignals subsequently translate into some form of appropriate buy orsell signal just prior to the major part of the relevant impulse wave.

It has been the objective of this book to show that technical analysisis an accurate and theoretically justifiable method of forecastingbehaviour in financial markets. Only putting theory into practice will,however, demonstrate the validity of the technique. Investors whofollow the rules outlined in this book should, after a little experience,find that it is entirely possible to embark on a long-term programmeof accumulating wealth.

Summary and conclusions 407

409

acceleration 119,171–71,173–74

A/D (Advance–Decline) line341–42

adaptation cycles 214, 220, 221,226, 234, 246

ADX (Average DirectionalIndicator) 341, 397, 398

affirmations 384alternation, rule of 298

see also Elliott wave principlealtruism 29American Civil War 218, 219,

220, 222, 223, 230, 233–34,244

amygdala 22,29anxiety 351, 358, 359 Australian All Ordinaries

187–89Average Directional Indicator

(ADX) 341averages (see moving averages)

avoidance compulsions 368–69,374, 375

base cycle 191,192, 193, 196,197, 198, 207, 208, 231, 233,262–63, 264

see also cyclesBateson, Gregory 19, 33, 41, 169bear see bull–bear cyclesbehaviour, human 361–76

see also personality typesbeliefs 12, 357–58

and goal setting 378–83groups 15–23

Berry, Brian 202Berry cycles 204–05, 209–12,

214–15, 222–26, 228–31,246, 247, 250

Bloom, Howard xibond markets 144–59, 257–69

UK 144–48, 153–56, 257–69US 149–51, 156–59

Index

boundaries 272, 306, 308–10,314–20

see also corrections; goldenratio

Brand, Stewart 271bull see bull–bear cyclesbull–bear cycles 78, 86–106,

288, 326and the economy 90–91and emotions 78, 80price pulse 101prices and sentiment 87–89,

92, 102reversal 97–99shocks 91–94, 102–03

Character see personality typesclosing positions 397, 400, 403collective consciousness 20complex corrections 295–96

see also Elliott wave principleconsciousness 19–21confidence 88, 327, 328, 331

see also sentimentconfirmation 326–47

direct, principle of 342–46and second price index

342–45conflict 29, 77corrections 302–24

alternation and 298boundaries 272, 306, 308–10,

314–20complex 295–96flat 294, 295, 298inverted 298running 294–95triangles 293, 296–97see also Elliott wave principle

crash 57–58, 59, 64–65, 84, 154,187–89, 190, 221, 243, 266,314

crisis cycles 214, 222, 223, 224,237–38, 243, 248–49, 250

see also cyclescrowd leader 26, 28–29, 37–38,

81–82crowds 4, 15–23, 73–84, 111,

118, 195financial markets 73–84and individuals 16, 30intelligence 19, 23life cycles 41–43see also groups

cyclesadaptation 214, 220, 221, 226,

234, 246and economy 201–15,

272and evolution 119–20,

213–14, 306base 191,192, 193, 196, 197,

198, 207, 208, 231, 233,262–63, 264

behavioural traits 195–97Berry 204–05, 209–12,

214–15, 222–26, 228–31,246, 247, 250

characteristics 191crisis 214, 222, 223, 224,

237–38, 243, 248–49, 250disruption 214, 233, 234, 235,

242, 250–51dyads 185–87and energy gaps 107–21,

185–89, 191, 195–95, 212, 231, 243, 266, 303,328

functions 191–92and industrial production

217–31innovation 214, 220, 229, 231,

247–48

410 Index

Juglar 204, 209–13, 220, 221,224–25, 227, 229, 233,234, 235, 236, 241, 246,247–49, 250

Kitchin 204, 205, 230, 251–55

Kondratyev 205, 206, 238–44

Kuznets 205recuperation 214, 229, 235,

236, 241, 242regeneration 214, 221, 229,

230, 236Strauss and Howe 204, 206,

222, 225, 228, 243–44terminal 191, 192, 193, 197,

198, 208, 223, 233translations 192–95trend 191, 192, 193, 196, 208,

220, 233triads 179–99, 204–05,

207–09, 219, 222–25see also bull–bear cycles; life

cycles; limit cycles

data see informationDawkins, Richard 18disruption cycles 214, 233, 234,

235, 242, 250–51direct confirmation, principle of

342–46directional indicator (DI)

340–41dollar 281, 283–84, 315–17,

319–24double retracements 292

see also Elliott wave principledoubling sequence 124Dow 174, 176–79, 181–86,

190–198, 254–55, 311, 314,342, 345

Dow theory 342–45

economic cycles 201–15, 272economic theory 206

problems with 61, 166,202–03

rational expectations 63egoElder, Dr Alexander 341Elliott, Ralph N 138, 286Elliott wave principle 5, 274–75,

284–300alternation, rule of 298basic wave patterns 287corrections 288, 290, 292,

294–98derived rules 289–90double retracements 292failures 291–92and impulse waves 290natural phenomenon 289and price pulse 287problems with 299–300variations 290

corrections 293–98fifth wave 291–93

emotions 76, 78, 331, 355–56bull–bear cycles 78influence 355–57see also fear

energy gaps 92–93, 107–21,185–87, 191, 194–95, 212,231, 243, 266, 303, 328

anticipating 327bridging 110–14confirmation 190and information flows

117–18and momentum 186–87

evolution 119–20, 213–14, 306economic 213–14

evolutionary metamorphosis114

extensions 291

Index 411

failures 291–92fear 99–100, 197, 331–32,

351–59, 373–74, 375feedback loops 36–38Fibonacci sequence 124–27First World War 220, 221, 234,

235five-wave patterns see Elliott wave

principle; price pulseflat corrections 294–95fractals 160, 217, 228

Gartley, H M 139Gilded Age 220, 234gilt-edged market 144–48,

153–56, 257–69goal-setting 378–86golden measure 128–29,

138–39golden ratio 123–39, 218,

282–84, 307–24boundaries 308–09, 315–17price targets 282–84retracement targets 310–313,

316support and resistance see

boundariessee also ratios

golden rectangles 130–31golden spiral 123–39Great Depression 209, 219, 222,

235, 236–37, 243groups 12, 15–23, 166–67, 195

see also crowds

habits 353–54health 387–89

Ichazo, Oscar 362, 367information 35–37, 112, 117–18,

134–35, 302–24crowds and 42–43

learning 107–09, 113, 319,323

and trends 303–04see also shocks

innovation cycles 214, 220, 229,231, 247–48

instinct 363–69and personality types 363–69

investment advisers 82–83

Jantsch Erich 33Juglar, Clement 202Juglar cycles 204, 209–13, 220,

221, 224–25, 227, 229, 233,234, 235, 236, 241, 246,247–49, 250

see also cycles

Kitchin, Joseph 202, 226,227–28

Kitchin Cycles 204, 205, 230,251–55

see also cyclesKondratyev, Nikolai 203,

238–44Kondratyev cycles 205, 206,

238–44see also cycles

Korean War 226Kuznets, Simon 202Kuznets cycles 205

see also cycles

Le Bon, Gustave 16, 17, 20, 39LeDoux Joseph 22learning 107, 113, 319, 323

curves 107–09see also information

life cycles 41–43, 102, 103,113–114

see also bull–bear cycles; cycleslimbic system 21, 22

412 Index

limit cycles 44–48, 102, 104,327, 331, 347

multiple 47and prices 87–89, 94–98and shocks 49–51

MACD (moving averageconvergence–divergence)338–39, 396, 397, 399, 400

Maclean, Paul 21, 342memes 18Milgram, Stanley 29Mills, Henry R 107–09mind 18–20, 33, 62, 134–35

see also subconsciousMissionary Awakening 220, 234,

235momentum 88, 100, 111, 119,

156–59, 169–74, 186–88,259, 328, 336–37, 340

and cycles 170–71, 259–60,263

motivation 11–13moving averages 337–38

see also MACD

Naranjo, Claudio 362neocortex 21, 22New Deal 225non-confirmation 172–74,

280–81, 327–29

open interest 328–35, 346, 347opening positions 396–97,

399–400overextended markets 327–28,

336

personality types 363–74avoidance compulsions

368–69basic motivations 367–68

characteristics 354–65and financial markets 372–75

price movements 54–60and crowd psychology 74–80mathematical relationships

140–52prediction methods 60see also bond market; price

pulseprice pulse 101, 153–61

see also Elliott wave principle;price movements; shocks

price reversals see reversalsprice targets 140–52

rational expectations 63ratios 123–39, 141–51, 282–84

1.618 formula 282–842.618 formula 141–51,

282–844.236 formula 282–84see also golden ratio

re-tests 96, 97–98, 111, 310, 327,332

recession 207, 208, 211, 212,214

rectangles, golden 130–32see also golden ratio

recuperation cycles 214, 229,235, 236, 241, 242

reductionism 10, 61regeneration cycles 214, 221,

229, 230, 236see also cycles

Relative Strength Index (RSI)339

see also Wells Wilderrelaxation techniques 390–91resistance see boundariesretracements 307–09, 310–13

double 292and pro-trend shocks 319, 320

Index 413

reversals 97–99see also boundaries

RSI (Relative Strength Index)339

see also Wells Wilderrunning corrections 294–95

see also Elliott wave principle

satiation 109, 113, 167, 212Schumacher, Fritz 12Schwager, Jack 1second price index 342–45Second World War 209, 234self-awareness 12, 20, 356self-organization 17–18, 33, 65,

136–37sentiment 62, 87–89, 94–96, 111,

328, 333shocks 48–50, 91–94, 96, 97, 98,

99, 102, 107–21, 302–24contra-trend 92–93hierarchical 315and limit cycles 48–50,

102–03pro-trend 92, 107–21, 319–20see also cycles; information;

price pulsespirals 49, 96, 98, 123–39, 327

golden 131–32stops 401Strauss and Howe cycle 204,

206, 222, 225, 228, 243–44stress 354–55subconscious 18, 352–53, 356support see boundariessystems theory 10

Tarde, Gabriel 107targets see price targetstechnical analysis 5, 66–68,

78–79, 326, 351

terminal cycle 191, 192, 193,197, 198, 208, 223, 233

threat responses 354–55three-wave patterns see cycles;

Elliott wave principle; pricepulse

trading systems 393–404closing positions 397, 400–01,

403design 395–96money management 399opening positions 396–97,

399–400, 403reversing positions 397spreading risk 398–99market map 398

translation, cycle 192–95Treasury bonds

UK 144–48, 153–56, 257–69

see also gilt-edged marketUS 149–51, 156–59

trend cycles 191, 192, 193, 196,208, 220, 233

see also cyclestriads 179–99, 204–05, 207–09,

219, 222–25see also cycles

triangles 293, 296–98corrections 296–97, 298fifth wave diagonal 293

triune brain 20–21

UK gilt-edged market 144–48,153–56, 257–69

US Treasury bond market149–51, 156–59

velocity see momentumVietnam War 209, 211, 226, volume 328–34

414 Index

waves see Elliott wave principle;price pulse

Wells Wilder, J 339

Yom Kippur War 227

zigzags 280, 293, 295, 298see also Elliott wave principle

Index 415

416


417


418


Tony Plummer - Forecasting Financial Markets The Psychology of Successful Investing 5th edition

Documents

Transcript of Tony Plummer - Forecasting Financial Markets The Psychology of Successful Investing 5th edition